Water plant pre-oxidation enhanced coagulation dosing system and intelligent regulation and control method

Abstract

The invention discloses a dosing system for pre-oxidation enhanced coagulation in a water plant and an intelligent regulation and control method, and relates to the technical field of pre-oxidation enhanced coagulation water treatment. The system comprises a raw water pump, a mixing tank, a tubular static mixer, a flocculation tank, a sedimentation tank, an industrial personal computer and multiple online detection instruments which are sequentially connected in series, so as to monitor the flow, conductivity, pH value, temperature, turbidity, ultraviolet absorbance and organic matter concentration of raw water. Key water quality indexes are monitored in real time through data collection and a pre-oxidation enhanced coagulation regulation and control system based on machine learning of an industrial personal computer, and intelligent and accurate prediction of the dosage of oxidants (potassium permanganate, permanganate composite agents and potassium ferrate) and coagulants is achieved; the effluent turbidity, the effluent ultraviolet absorbance and the effluent organic matter concentration of the sedimentation tank are fed back in real time, the adding amount of the oxidizing agent and the coagulant is dynamically adjusted through a regulation and control algorithm, the effluent quality is guaranteed, and medicine consumption is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of water treatment, discloses a dosing system and an intelligent control method for pre-oxidation enhanced coagulation in a water plant, and particularly relates to the technical field of pre-oxidation enhanced coagulation water treatment. Background Art

[0002] Under the background of the implementation of the new drinking water hygiene standard (GB5749-2022), algae-derived complex organic pollution (including algae cells, algae odor-causing organic matter and algae toxins) caused by the massive reproduction and death of algae during the high algae period seriously threatens the water safety of residents and brings many challenges to water supply companies. Coagulation is one of the most important unit technologies in the drinking water treatment process, mainly used to remove colloids and suspended particles. However, with the increasing severity of organic complex pollution caused by natural organic matter in natural water bodies and harmful algal blooms in summer, as well as the stricter water quality standards, conventional coagulation technology can no longer meet the current water quality safety requirements.

[0003] Chemical pre-oxidation enhanced coagulation is the core algae removal technology adopted by most domestic water plants during the high algae period. Its principle is to use pre-oxidation to inactivate algae cells or destroy their activity and floating mechanism, making them easier to be removed by coagulation precipitation or flotation. Common pre-oxidants include liquid chlorine, chlorine dioxide, sodium hypochlorite, ozone, potassium permanganate, permanganate compound agent, potassium ferrate, etc. Among them, liquid chlorine, chlorine dioxide, sodium hypochlorite, ozone, etc. are strong oxidizing algae removal agents; potassium permanganate, permanganate compound agent, potassium ferrate are mild oxidizing algae removal agents.

[0004] Pre-chlorination can enhance the removal of algae, but it cannot eliminate the smell caused by algae, but may make it more unpleasant. During the pre-chlorination process, chlorine reacts with high concentrations of organic matter in raw water to produce a large number of highly toxic disinfection by-products that are harmful to the human body. Chlorine dioxide is a strong oxidant with an oxidation capacity between chlorine and ozone. It also has an algaecidal effect and can effectively control the color, odor and taste of water. Chlorine dioxide can damage the structure of algae cells, leading to the release of organic matter in algae cells. In addition, by-products produced by chlorine dioxide pre-oxidation, such as chlorite, chlorate and formaldehyde, can damage human red blood cells, so the dosage must also be strictly controlled to avoid secondary pollution. Sodium hypochlorite mainly reflects its oxidizing properties in the form of available chlorine, so it will generate highly carcinogenic disinfection by-products and chlorates. Although chlorine dioxide is also measured in available chlorine, its oxidizing properties are mainly reflected in atomic oxygen, which has a stronger inactivation effect on algae. Due to production safety control reasons, liquid chlorine is gradually replaced by sodium hypochlorite and chlorine dioxide. In addition, liquid chlorine and sodium hypochlorite are very likely to react with organic matter in water to generate disinfection by-products during the oxidation process, and chlorine dioxide will generate inorganic chlorate and chlorite and other by-products, so the dosage must be strictly controlled. In addition, liquid chlorine, sodium hypochlorite and chlorine dioxide will destroy the cell wall of algae during the algae killing process, and the broken algae and its extracellular secretions will cause the filter pool to be blocked or the ultrafiltration membrane pores to be blocked.

[0005] As a strong oxidant with a high redox potential, ozone's algae removal mechanism is mainly characterized by strong oxidizing properties. During the ozone pre-oxidation process, it can quickly desorb organic matter outside algae cells, inhibit the activity of algae cells, and indirectly produce hydroxyl free radicals with stronger oxidizing properties, aggravating the oxidative damage to algae cells. Ozone pre-oxidation will also produce soluble organic nitrogen and small molecular disinfection by-product precursors. Excessive doses of ozone pre-oxidation will also change the structure of organic matter outside algae cells in a short period of time, increasing the amount of biodegradable organic matter in the water.

[0006] Potassium permanganate can effectively inhibit algae, and its pre-oxidation enhanced coagulation effect is mainly reflected in two aspects: first, potassium permanganate can destroy the organic gel layer through oxidation, and under the stimulation of oxidants, algal cells will secrete more algal extracellular organic matter to enhance the coagulation effect; on the other hand, potassium permanganate can produce new ecological hydrated manganese dioxide after reacting with various substances in water. Hydrated manganese dioxide can be adsorbed on the surface of algae, change the surface characteristics of algae, increase the specific gravity of algal flocs, thereby improving the sedimentation performance of algae and improving the filtration performance. Potassium permanganate can significantly improve the removal rate of algae in water (from 70% to 80%). Low concentrations of potassium permanganate can maintain the integrity of algal cells and avoid the release of a large amount of algal extracellular organic matter. However, potassium permanganate has a heavier color. Excessive addition will increase the chromaticity of the effluent and increase the manganese content in the water. Its dosage should be strictly controlled.

[0007] The oxidizing capacity of the permanganate compound agent is lower than that of ozone, but higher than that of sodium hypochlorite and chlorine dioxide. During pre-oxidation, the permanganate compound agent kills algal cells by taking away electrons from algal cell proteins. The new ecological divalent manganese (δ-MnO2) produced during the oxidation process can cross-link with the extracellular organic matter of the algae, increasing the density and sedimentability of the algal cells.

[0008] In summary, although chemical pre-oxidation can effectively enhance the effect of coagulation and algae removal, it will also cause algal cell rupture, the release of extracellular and intracellular organic matter, and significantly increase the generation potential of disinfection by-products. Pre-chlorination, chlorine dioxide pre-oxidation and pre-ozonation cause serious damage to the structure of algal cells. Therefore, the pre-oxidation treatment process needs to maintain a balance between improving the removal rate of pollutants and reducing the potential water quality risks caused by excessive oxidation in order to effectively remove algae pollution. In order to minimize the rupture of algal cells and the release of algae organic matter, the risk of oxidation can be reduced by controlling the dosage of oxidants, while ensuring that the algal cells are not ruptured, inhibiting their activity and achieving the purpose of enhancing algae removal. Therefore, the selection of oxidants and coagulants and the dosage of the two are the key to controlling algae pollution.

[0009] Surface water plants still have the problem of poor coagulation effect under the condition of seasonal water source water quality changes. This is mainly because the addition of oxidants (such as potassium permanganate, permanganate compound agents, potassium ferrate) and coagulants used in water plants for algae removal is random and blind, and lacks systematic, scientific technology and theoretical support. With the seasonal changes in the quality of raw water from surface water plants, the amount of coagulants and oxidants added in different periods is different, which is mainly affected by water temperature, turbidity, organic matter content and algae. These factors will directly affect the effect of pre-oxidation and coagulation in the water treatment process. Therefore, it is urgent to achieve refined, accurate and intelligent control of the amount of oxidants and coagulants added while ensuring that the effluent water quality meets the standards. Summary of the invention

[0010] In order to solve the above technical problems, the present invention provides a dosing system and method for pre-oxidation enhanced coagulation in a water plant, specifically including a pre-oxidation enhanced coagulation dosing system, and a control method for the combined dosing of oxidants and coagulants based on data-driven and knowledge fusion, which not only reduces the drug consumption to a certain extent, but also can remove organic matter to the maximum extent to achieve the depollution effect of enhanced coagulation, thereby providing a guarantee for the safety of drinking water supply.

[0011] The technical solution adopted by the system of the present invention is: a dosing system for pre-oxidation and enhanced coagulation in a water plant, characterized in that:

[0012] It includes a raw water pump (1), a mixing tank (2), a tubular static mixer (3), a flocculation tank (4), a sedimentation tank (5), a flow meter (6), an online conductivity detector (7), an online pH detector (8), an online water temperature monitor (9), a No. 1 online turbidity meter (10), a No. 1 online ultraviolet detector (11), a No. 1 online organic matter analyzer (12), a No. 2 online turbidity meter (13), a No. 2 online ultraviolet detector (14), a No. 2 online organic matter analyzer (15), an oxidant tank (16), a coagulant tank (17), a No. 1 metering pump (18), a No. 2 metering pump (19), a No. 1 variable frequency speed controller (20), a No. 2 variable frequency speed controller (21), and an industrial computer (22);

[0013] The mixing tank (2), the tubular static mixer (3), the flocculation tank (4), and the sedimentation tank (5) are sequentially connected in series by pipelines to form a water treatment structure; raw water is sent into the mixing tank (2) through the raw water pump (1);

[0014] The flocculation tank (4) may be a mechanical reaction flocculation tank or a hydraulic reaction flocculation tank, such as a baffle reaction flocculation tank, a perforated plate reaction flocculation tank, a corrugated plate reaction flocculation tank, etc.; the sedimentation tank (5) may be an inclined plate sedimentation tank, an inclined tube sedimentation tank or a horizontal flow sedimentation tank;

[0015] The flow meter (6), the online conductivity detector (7), the online pH detector (8), the online water temperature monitor (9), the No. 1 online turbidity meter (10), the No. 1 online ultraviolet detector (11), and the No. 1 online organic matter analyzer (12) are arranged on the raw water inlet pipe, and the No. 2 online turbidity meter (13), the No. 2 online ultraviolet detector (14), and the No. 2 online organic matter analyzer (15) are arranged on the sedimentation tank outlet pipe;

[0016] The oxidant tank (16) adds oxidant to the mixing tank (2) through a No. 1 metering pump (18), and the coagulant tank (17) adds coagulant to the tubular static mixer (3) through a No. 2 metering pump (19); the oxidant can be a mild oxidant such as potassium permanganate, a permanganate compound agent, and potassium ferrate, and the coagulant is an aluminum salt coagulant (polyaluminum chloride and aluminum sulfate, etc.); preferably, the reaction time of the oxidant in the mixing tank (2) is 5 minutes.

[0017] The first variable frequency speed regulating controller (20) is electrically connected to the first metering pump (18) for controlling the dosage of the oxidant, and the second variable frequency speed regulating controller (21) is electrically connected to the second metering pump (19) for controlling the dosage of the coagulant; the industrial computer (22) is electrically connected to the flow meter (6), the online conductivity detector (7), the online pH detector (8), the online water temperature monitor (9), the first online turbidity meter (10), the first online ultraviolet detector (11), the first organic matter online analyzer (12), the second online turbidity meter (13), the second online ultraviolet detector (14), the second organic matter online analyzer (15), the first variable frequency speed regulating controller (20) and the second variable frequency speed regulating controller (21); the industrial computer (22) has a trained machine learning model.

[0018] The flow meter (6) monitors the raw water flow in real time, the online conductivity detector (7) monitors the raw water conductivity in real time, the online pH detector (8) monitors the raw water pH in real time, the online water temperature monitor (9) monitors the water temperature in real time, the No. 1 online turbidity meter (10) monitors the raw water turbidity in real time, and the No. 1 online ultraviolet detector (11) monitors the raw water ultraviolet absorbance (UV 254 ), No. 1 organic matter online analyzer (12) monitors the raw water organic matter concentration (COD Mn ), and transmits the above-collected data to the industrial computer (22), which automatically adjusts the dosage of the oxidant and the coagulant according to the result predicted by the machine learning model, that is, under the instruction of the industrial computer (22), changes the rotation speed of the first metering pump (18) and the second metering pump (19) by controlling the first variable frequency speed control controller (20) and the second variable frequency speed control controller (21).

[0019] The technical solution adopted by the system of the present invention is: a dosing method for pre-oxidation and enhanced coagulation in a water plant based on machine learning, characterized in that it includes the following steps:

[0020] Step 1: Data collection and model training on industrial computers; Collect test data, including determining the optimal dosage of the combination of oxidant and coagulant under different raw water quality conditions (conductivity, pH, water temperature, turbidity, UV absorbance, organic matter concentration) through experiments. The optimal dosage is the removal efficiency of the target pollutants (turbidity removal rate, UV absorbance, organic matter concentration). 254 Removal rate, COD Mn The machine learning model was used to train the test data set, establish the relationship between different raw water quality conditions and the optimal oxidant and coagulant combination, and obtain the regression prediction model between raw water quality and the oxidant and coagulant dosage;

[0021] Step 2: Model integration and industrial control system deployment: Integrate the model trained in step 1 into the industrial control computer (22) system to ensure that the model can receive real-time data input and establish an automated monitoring and control system;

[0022] Step 3: Real-time monitoring and data input; using a flow meter (6), an online conductivity detector (7), an online pH detector (8), an online water temperature monitor (9), a No. 1 online turbidity meter (10), a No. 1 online ultraviolet detector (11), and a No. 1 online organic matter analyzer (12) to monitor the flow rate, conductivity, pH, water temperature, turbidity, ultraviolet absorbance, and organic matter concentration of the raw water in real time, and inputting these real-time data into the model of the industrial computer (22) system to predict the dosage of the oxidant and coagulant;

[0023] Step 4: dosage adjustment and optimization; according to the model prediction results and the raw water flow rate, the industrial computer (22) system automatically adjusts the dosage of the oxidant and coagulant by controlling the No. 1 variable frequency speed controller (20) and the No. 2 variable frequency speed controller (21) to change the rotation speed of the No. 1 metering pump (18) and the No. 2 metering pump (19), so as to achieve accurate dosage of the appropriate pre-oxidation enhanced coagulation dosage;

[0024] Step 5: Continuous monitoring and parameter adjustment; Use the No. 2 online turbidity meter (13), the No. 2 online UV detector (14), and the No. 2 online organic matter analyzer (15) to continuously monitor the turbidity of the sedimentation tank effluent, the effluent UV absorbance, and the changes in the effluent organic matter concentration (the operating condition machine can obtain the turbidity removal rate, UV 254 Removal rate, COD Mn Removal rate) is used to evaluate the water treatment situation. As the raw water quality conditions (conductivity, pH, water temperature, turbidity, ultraviolet absorbance, organic matter concentration) are fed back according to monitoring results and model predictions, the model parameters are adjusted in a timely manner through intelligent control methods to achieve automatic control effects.

[0025] In the present invention, the machine learning model is a radial basis function (RBF) neural network model. The RBF neural network can map the input data to a high-dimensional space through a radial basis function, and can capture complex nonlinear relationships in the space. Therefore, it has a strong ability to deal with nonlinear problems and is suitable for most scenarios that require nonlinear fitting. Compared with other types of neural networks, the training process of the RBF neural network is usually simpler. The training speed is relatively fast, especially when dealing with smaller data sets. It has strong generalization ability and can maintain good prediction accuracy with a small number of training samples. Especially when dealing with tasks with more noise or uneven data distribution, the RBF neural network can avoid overfitting and show good robustness. Compared with deep neural networks, RBF neural networks usually require less hyperparameter adjustment.

[0026] In the present invention, the RBF neural network model is designed to predict the dosage of oxidant and coagulant, wherein the conductivity, pH, water temperature, turbidity, ultraviolet absorbance, and organic matter concentration of the raw water are used as input features, and the dosage of oxidant and coagulant is used as output features. A total of 53 sample data sets are collected, of which 80% of the samples are used for RBF neural network model training, and the remaining 20% ​​are used for model performance testing. The training and test data sets are re-divided ten times, and different sample allocations are performed to ensure the reliability and robustness of the model. MAE (mean absolute error), MBE (mean deviation error), RMSE (root mean square error) and R 2 (Coefficient of determination) Four statistical parameters were used to evaluate the prediction accuracy of RBF neural networks.

[0027] The beneficial effects of the present invention are:

[0028] The present invention establishes a dosing system and method for pre-oxidation enhanced coagulation in a water plant, which can realize automatic control of the pre-oxidation enhanced coagulation process and ensure that the dosage of the agent is in the optimal range. The system adopts an innovative machine learning-driven pre-oxidation enhanced coagulation control strategy, which monitors the key raw water quality indicators (conductivity, pH, water temperature, turbidity, ultraviolet absorbance, organic matter concentration) of the raw water in real time, and tracks the changes in the turbidity, ultraviolet absorbance and organic matter concentration of the effluent after pre-oxidation enhanced coagulation treatment in real time, and uses a radial basis function (RBF) neural network and an intelligent control algorithm to accurately predict and automatically adjust the dosage of the oxidant and coagulant. This intelligent solution significantly improves the stability and safety of water quality while saving the cost of agents. By reducing manual intervention, the system enhances the responsiveness of the water plant to the addition of chemical agents. Therefore, the present invention is suitable for the optimization of the coagulation process of a water plant, and is suitable for improving the quality and efficiency of existing water plants and for application in new water plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the process flow and control system of the invention. 1-raw water pump; 2-mixing tank; 3-tubular static mixer; 4-flocculation tank; 5-sedimentation tank; 6-flow meter; 7-online conductivity detector; 8-online pH detector; 9-online water temperature monitor; 10-No. 1 online turbidity meter; 11-No. 1 online ultraviolet detector; 12-No. 1 organic matter online analyzer; 13-No. 2 online turbidity meter; 14-No. 2 online ultraviolet detector; 15-No. 2 organic matter online analyzer; 16-oxidant medicine tank; 17-coagulant medicine tank; 18-No. 1 metering pump; 19-No. 2 metering pump; 20-No. 1 variable frequency speed controller; 21-No. 2 variable frequency speed controller; 22-industrial computer. DETAILED DESCRIPTION

[0030] The present invention is described in detail through the following examples in combination with the accompanying drawings. This embodiment is only a part of the examples of the present invention, rather than all the embodiments.

[0031] Implementation Example 1:

[0032] The present invention establishes a dosing system for water plant pre-oxidation enhanced coagulation based on machine learning model. Through the monitoring of key raw water quality indicators (conductivity, pH, water temperature, turbidity, ultraviolet absorbance, organic matter concentration) and effluent water quality indicators (turbidity, ultraviolet absorbance, organic matter concentration), accurate prediction and automatic adjustment of oxidant and coagulant dosage are achieved, thereby achieving the purpose of enhanced coagulation and reducing the subsequent process operation load.

[0033] Please see Figure 1 The present invention provides a dosing system based on pre-oxidation enhanced coagulation in a water plant, comprising a raw water pump 1, a mixing tank 2, a tubular static mixer 3, a flocculation tank 4, a sedimentation tank 5, a flow meter 6, an online conductivity detector 7, an online pH detector 8, an online water temperature monitor 9, a No. 1 online turbidity meter 10, a No. 1 online ultraviolet detector 11, a No. 1 online organic matter analyzer 12, a No. 2 online turbidity meter 13, a No. 2 online ultraviolet detector 14, a No. 2 online organic matter analyzer 15, an oxidant medicine tank 16, a coagulant medicine tank 17, a No. 1 metering pump 18, and a No. 2 Metering pump 19, No. 1 frequency conversion speed controller 20, No. 2 frequency conversion speed controller 21, industrial computer 22; slightly polluted raw water is sent into mixing tank 2 and tubular static mixer 3 through raw water pump 1; flow meter 6, online conductivity detector 7, online pH detector 8, online water temperature monitor 9, No. 1 online turbidity meter 10, No. 1 online ultraviolet detector 11, No. 1 online organic matter analyzer 12 are arranged on the raw water inlet pipeline, and No. 2 online turbidity meter 13, No. 2 online ultraviolet detector 14, No. 2 online organic matter analyzer 15 are arranged on the sedimentation tank outlet pipeline;

[0034] The oxidant tank 16 and the coagulant tank 17 respectively add oxidant and coagulant to the mixing tank 2 and the tubular static mixer 3 through the No. 1 metering pump 18 and the No. 2 metering pump 19, and the No. 1 variable frequency speed controller 20 and the No. 2 variable frequency speed controller 21 are electrically connected to the No. 1 metering pump 18 and the No. 2 metering pump 19 respectively; the flow meter 6 monitors the raw water flow in real time, the online conductivity detector 7 monitors the raw water conductivity in real time, the online pH detector 8 monitors the raw water pH in real time, the online water temperature monitor 9 monitors the raw water temperature, the No. 1 online turbidity meter 10 monitors the raw water turbidity in real time, the No. 1 online ultraviolet detector 11 monitors the raw water ultraviolet absorbance in real time, and the No. 1 organic matter online analyzer 12 monitors the raw water organic matter concentration in real time, and transmits the collected data to the industrial computer 22, and the industrial computer 22 system automatically adjusts the dosage of the oxidant and coagulant according to the results predicted by the machine learning model.

[0035] The specific working process of the system of the present invention is to collect test data, including determining the optimal dosage of the combination of oxidant and coagulant under different raw water quality conditions (conductivity, pH, water temperature, turbidity, UV absorbance, organic matter concentration) through experiments. The optimal dosage is the removal efficiency of the target pollutants (turbidity removal rate, UV 254 Removal rate, COD Mn The method is to use a machine learning model to train the test data set, establish the relationship between different raw water quality conditions and the optimal oxidant and coagulant dosage, and realize the regression prediction of raw water quality and oxidant and coagulant dosage; integrate the trained model into the industrial computer 22 system to ensure that the model can receive real-time data input and establish an automated monitoring and control system; use a flow meter 6, an online conductivity detector 7, an online pH detector 8, an online water temperature monitor 9, a No. 1 online turbidity meter 10, a No. 1 online ultraviolet detector 11, and a No. 1 organic matter online analyzer 12 to monitor the flow, conductivity, pH, water temperature, turbidity, ultraviolet absorbance, and organic matter concentration of the raw water in real time, and input these real-time data into the industrial computer 2 2 system model to predict the dosage of oxidant and coagulant; according to the model prediction results and the raw water flow rate, the industrial computer 22 system changes the rotation speed of the No. 1 metering pump 18 and the No. 2 metering pump 19 by controlling the No. 1 variable frequency speed controller 20 and the No. 2 variable frequency speed controller 21 to automatically adjust the dosage of oxidant and coagulant, so as to achieve precise control and optimize the pre-oxidation enhanced coagulation treatment effect; use the No. 2 online turbidity meter 13, the No. 2 online ultraviolet detector 14, and the No. 2 organic matter online analyzer 15 to continuously monitor the changes in the turbidity of the sedimentation tank effluent, the effluent ultraviolet absorbance, and the effluent organic matter concentration to evaluate the performance of the system; according to the monitoring results and the feedback of the model prediction, adjust the model parameters in time to ensure the accuracy of the prediction and the suitability of the dosage of the oxidant and coagulant.

[0036] The machine learning model used in the industrial computer 22 is the radial basis function (RBF) neural network model, and the following indicators are used to evaluate the model performance: mean absolute error (MAE), mean bias error (MBE), root mean square error (RMSE), and determination coefficient (R 2 ). The performance of the RBF neural network on the test set is as follows: the prediction performance of oxidants is: MAE is 0.0542, MBE is -0.0061, RMSE is 0.0709, R 2 The prediction performance of coagulant is: MAE is 7.196, MBE is -1.0596, RMSE is 9.0522, R 2is 0.747. Compared with other machine learning models, RBF neural network has significantly improved prediction accuracy and has better prediction performance. The effectiveness of RBF neural network application has been proven, and it can accurately predict the dosage of oxidant and coagulant through the changes in raw water quality.

[0037] Implementation Example 2:

[0038] The dosing system for water plant pre-oxidation enhanced coagulation based on a machine learning model established by the present invention can automatically adjust the dosage of oxidant and coagulant according to the seasonal changes in the raw water quality.

[0039] During the high algae period, the conductivity of the slightly polluted raw water of a surface water plant at a certain moment was 562μS / cm, pH was 7.8, water temperature was 29.3℃, turbidity was 17.4NTU, UV 254 0.112cm -1 , COD Mn The water body is highly polluted by organic matter. According to the prediction of the preset RBF neural network model, the dosage of oxidant (potassium permanganate) is 0.8 mg / L, and the dosage of coagulant (polyaluminium chloride) is 0.13 mM (in terms of Al). At this time, the turbidity of the sedimentation tank effluent is 0.86 NTU, and UV 254 0.060, COD Mn It is 3.47mg / L.

[0040] During the low temperature and low turbidity period, the conductivity at a certain moment was 771μS / cm, pH was 8.2, water temperature was 6.4℃, turbidity was 7.05NTU, UV 254 0.058cm -1 , COD Mn The water body is highly polluted and the algae stage is low, but the lower temperature and turbidity will seriously affect the flocculation effect. According to the prediction of the preset RBF neural network model, the dosage of potassium permanganate is 0.5mg / L and the dosage of polyaluminium chloride is 0.12mM (in terms of Al). At this time, the turbidity of the sedimentation tank effluent is 0.86NTU, UV 254 0.039, COD Mn It is 2.00 mg / L, and the removal effect of turbidity and organic pollutants is good, and the residual turbidity and organic matter levels are low.

[0041] In spring, at a certain moment, the conductivity was 826μS / cm, the pH was 8.0, the water temperature was 15.5℃, the turbidity was 16.4NTU, and the UV 254 0.047cm -1 , COD MnThe water pollution degree is the lightest compared with other periods, and the temperature and turbidity do not affect the flocculation effect during this period. According to the prediction of the preset RBF neural network model, the potassium permanganate dosage is the lowest at this time, which is 0.3 mg / L, and the polyaluminium chloride dosage is also the lowest, which is 0.08 mM (in terms of Al). At this time, the turbidity of the sedimentation tank effluent is 0.64 NTU, and UV 254 0.033, COD Mn It is 1.64 mg / L, with the best removal effect on turbidity and organic pollutants, and the lowest levels of residual turbidity and organic pollutants.

[0042] The same water treatment plant at the same location mentioned above can automatically adjust to the optimal matching dosage of oxidant and coagulant as the water quality changes by using the system of the present invention.

Claims

1. A dosing system for pre-oxidation and enhanced coagulation in a water plant, characterized by: It includes a raw water pump (1), a mixing tank (2), a tubular static mixer (3), a flocculation tank (4), a sedimentation tank (5), a flow meter (6), an online conductivity detector (7), an online pH detector (8), an online water temperature monitor (9), a No. 1 online turbidity meter (10), a No. 1 online ultraviolet detector (11), a No. 1 online organic matter analyzer (12), a No. 2 online turbidity meter (13), a No. 2 online ultraviolet detector (14), a No. 2 online organic matter analyzer (15), an oxidant tank (16), a coagulant tank (17), a No. 1 metering pump (18), a No. 2 metering pump (19), a No. 1 variable frequency speed controller (20), a No. 2 variable frequency speed controller (21), and an industrial computer (22); The mixing tank (2), the tubular static mixer (3), the flocculation tank (4), and the sedimentation tank (5) are connected in series in sequence by pipelines to form a water treatment structure; raw water is sent into the mixing tank (2) through the raw water pump (1); The flow meter (6), the online conductivity detector (7), the online pH detector (8), the online water temperature monitor (9), the No. 1 online turbidity meter (10), the No. 1 online ultraviolet detector (11), and the No. 1 online organic matter analyzer (12) are arranged on the raw water inlet pipe, and the No. 2 online turbidity meter (13), the No. 2 online ultraviolet detector (14), and the No. 2 online organic matter analyzer (15) are arranged on the sedimentation tank outlet pipe; The oxidant tank (16) adds oxidant into the mixing tank (2) through a No. 1 metering pump (18), and the coagulant tank (17) adds coagulant into the tubular static mixer (3) through a No. 2 metering pump (19); The first variable frequency speed regulating controller (20) is electrically connected to the first metering pump (18) for controlling the dosage of the oxidant, and the second variable frequency speed regulating controller (21) is electrically connected to the second metering pump (19) for controlling the dosage of the coagulant; the industrial computer (22) is electrically connected to the flow meter (6), the online conductivity detector (7), the online pH detector (8), the online water temperature monitor (9), the first online turbidity meter (10), the first online ultraviolet detector (11), the first organic matter online analyzer (12), the second online turbidity meter (13), the second online ultraviolet detector (14), the second organic matter online analyzer (15), the first variable frequency speed regulating controller (20) and the second variable frequency speed regulating controller (21); the industrial computer (22) has a trained machine learning model.

2. A dosing system for pre-oxidation and enhanced coagulation in a water plant according to claim 1, characterized in that: The flocculation tank (4) is a mechanical reaction flocculation tank or a hydraulic reaction flocculation tank, such as a baffle reaction flocculation tank, a perforated plate reaction flocculation tank, a corrugated plate reaction flocculation tank, etc.; the sedimentation tank (5) is an inclined plate sedimentation tank, an inclined tube sedimentation tank or a horizontal flow sedimentation tank.

3. A dosing system for pre-oxidation and enhanced coagulation in a water plant according to claim 1, characterized in that: The flow meter (6) monitors the raw water flow in real time, the online conductivity detector (7) monitors the raw water conductivity in real time, the online pH detector (8) monitors the raw water pH in real time, the online water temperature monitor (9) monitors the water temperature in real time, the No. 1 online turbidity meter (10) monitors the raw water turbidity in real time, and the No. 1 online ultraviolet detector (11) monitors the raw water ultraviolet absorbance (UV 254 ), No. 1 organic matter online analyzer (12) monitors the raw water organic matter concentration (COD Mn ), and transmits the above-collected data to the industrial computer (22), which automatically adjusts the dosage of the oxidant and the coagulant according to the result predicted by the machine learning model, that is, under the instruction of the industrial computer (22), changes the rotation speed of the first metering pump (18) and the second metering pump (19) by controlling the first variable frequency speed control controller (20) and the second variable frequency speed control controller (21).

4. A dosing method for pre-oxidation enhanced coagulation using the dosing system according to any one of claims 1 to 3, characterized in that: The following steps are involved: Step 1: Industrial computer data collection and model training; collect test data, including determining the optimal dosage of the combination of oxidants and coagulants under different raw water quality conditions through experiments. The optimal dosage is the dosage with the highest removal efficiency of the target pollutants and the most economical dosage of the oxidants and coagulants used. Use the machine learning model to train the test data set, establish the relationship between different raw water quality conditions and the optimal combination of oxidants and coagulants, and obtain the regression prediction model of raw water quality and the dosage of oxidants and coagulants; the raw water quality conditions include conductivity, pH, water temperature, turbidity, ultraviolet absorbance and organic matter concentration, and the removal efficiency of the target pollutants is turbidity removal rate, UV 254 Removal rate and COD Mn Removal rate; Step 2: Model integration and industrial control system deployment: Integrate the model trained in step 1 into the industrial control computer (22) system to ensure that the model can receive real-time data input and establish an automated monitoring and control system; Step 3: Real-time monitoring and data input; using a flow meter (6), an online conductivity detector (7), an online pH detector (8), an online water temperature monitor (9), a No. 1 online turbidity meter (10), a No. 1 online ultraviolet detector (11), and a No. 1 online organic matter analyzer (12) to monitor the flow rate, conductivity, pH, water temperature, turbidity, ultraviolet absorbance, and organic matter concentration of the raw water in real time, and inputting these real-time data into the model of the industrial computer (22) system to predict the dosage of the oxidant and coagulant; Step 4: dosage adjustment and optimization; according to the model prediction results and the raw water flow rate, the industrial computer (22) system automatically adjusts the dosage of the oxidant and coagulant by controlling the No. 1 variable frequency speed controller (20) and the No. 2 variable frequency speed controller (21) to change the rotation speed of the No. 1 metering pump (18) and the No. 2 metering pump (19), so as to achieve accurate dosage of the appropriate pre-oxidation enhanced coagulation dosage; Step 5: Continuous monitoring and parameter adjustment; Use the No. 2 online turbidity meter (13), the No. 2 online UV detector (14), and the No. 2 online organic matter analyzer (15) to continuously monitor the turbidity of the sedimentation tank effluent, the effluent UV absorbance, and the changes in the effluent organic matter concentration (the operating condition machine can obtain the turbidity removal rate, UV 254 Removal rate, COD Mn Removal rate) is used to evaluate the water treatment situation. As the raw water quality conditions are fed back based on monitoring results and model predictions, the model parameters are adjusted in a timely manner through intelligent control methods to achieve automatic control effects.

5. The method according to claim 4, characterized in that The machine learning model described in step (1) is a radial basis function (RBF) neural network model. The RBF neural network can map input data to a high-dimensional space through a radial basis function, and can capture complex nonlinear relationships in the space.

6. The method according to claim 4, characterized in that The RBF neural network model is designed to predict the dosage of oxidant and coagulant, where the conductivity, pH, water temperature, turbidity, ultraviolet absorbance, and organic matter concentration of raw water are used as input features, and the dosage of oxidant and coagulant is used as output features. A total of 53 sample data sets are collected, of which 80% of the samples are used for RBF neural network model training, and the remaining 20% ​​are used for model performance testing; the training and test data sets are re-divided ten times, and different sample allocations are performed to ensure the reliability and robustness of the model; MAE (mean absolute error), MBE (mean deviation error), RMSE (root mean square error) and R 2 (Coefficient of determination) Four statistical parameters were used to evaluate the prediction accuracy of RBF neural networks.

Images