A device and control method for the harmless treatment of chlorophenol compounds in wastewater

Abstract

This invention discloses a device and control method for the harmless treatment of chlorophenol compounds in wastewater, comprising a wastewater adsorption system, a thermal desorption system, a catalytic combustion system, and an intelligent control system. The main component of the wastewater adsorption system is an activated carbon adsorption tower, and the chlorine content in the water is measured using a water quality analyzer and a flow meter. The main component of the thermal desorption system is a desorption rotary kiln, and the desorption heat is calculated using a desorption gas flow meter and a thermometer. The main component of the catalytic combustion system is a catalytic burner, and the catalyst bed temperature and oxidant dosage are calculated using a thermometer, a flow meter, and an oxygen analyzer. All four calculated values ​​are simultaneously fed into the intelligent control system. After data dimensionality reduction and normalization, the data is fed into a neural network model to calculate a predicted value for dioxins. The predicted value is then fed into a logic control module to compare with the set value. Finally, the valve control module automatically adjusts the wastewater flow rate, desorption gas flow rate, and oxidant flow rate.

Description

Technical Field

[0001] This invention relates to the fields of industrial wastewater treatment and environmental management, as well as related automation control, and particularly to a device and control method for the harmless treatment of chlorophenol compounds in wastewater. Background Technology

[0002] Coal chemical processes mainly consist of coal coking, coal gasification, coal liquefaction, and tar chemical processes. During the operation of coal chemical plants, coking wastewater, gasification wastewater, and liquefaction wastewater are generated depending on the process. Highly toxic and difficult-to-degrade chlorophenolic compounds in these wastewaters will eventually enter the wastewater through processes such as washing and condensation. Improper subsequent treatment can easily cause secondary pollution.

[0003] In the treatment of coal chemical wastewater, activated carbon adsorption is a commonly used and effective method. Activated carbon, due to its high specific surface area, abundant surface functional groups, and excellent regenerability, is a good adsorbent in wastewater treatment, capable of adsorbing and enriching chlorophenol compounds in wastewater. Studies have shown that while activated carbon is generally used as an adsorbent for chlorophenol-containing wastewater, the presence of chlorophenols during the treatment or regeneration of spent activated carbon can generate highly toxic dioxins in the presence of oxygen.

[0004] Saturated activated carbon regeneration typically employs a thermal regeneration method, where the high-temperature flue gas from natural gas combustion heats the saturated activated carbon, causing the adsorbed volatile substances to desorb and restoring the activated carbon's pore structure. For saturated activated carbon regeneration processes with high chlorophenol content, incomplete combustion of the generated flue gas in the secondary combustion chamber, or the absence of a dioxin-specific catalyst in the catalytic combustion process, both contribute to excessive dioxin emissions. Therefore, using a dioxin-specific catalyst for secondary catalytic combustion is essential.

[0005] Dioxins are a group of polychlorinated oxygen-containing tricyclic aromatic hydrocarbons. Their formation mechanism is complex and related to many factors such as chlorine source, oxygen source, and temperature. The existing method for detecting dioxin concentration in the secondary treatment of waste activated carbon is high-resolution chromatography-high-resolution double-focusing magnetic mass spectrometry. However, the detection cycle is long and the cost is high. Therefore, it is necessary to conduct soft measurement of dioxin concentration.

[0006] To better perform soft measurement, the above factors need to be considered comprehensively. However, since the generation mechanism of dioxins is not yet clear and it is a multivariate control process, intelligent control systems based on neural network models can effectively solve the problem of soft measurement. Summary of the Invention

[0007] The purpose of this invention is to provide a safe, environmentally friendly, and highly automated device and intelligent control method for the harmless treatment of chlorophenol compounds in wastewater.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A device for the harmless treatment of chlorophenol compounds in wastewater, characterized in that it comprises a wastewater adsorption system, a thermal desorption system, a catalytic combustion system, and an intelligent control system, wherein:

[0010] The wastewater adsorption system includes chlorophenol-containing wastewater from upstream, an activated carbon adsorption tower, purified water going downstream, fresh activated carbon from upstream, water transported from the public utility network, a circulating water pump, and a water transport tank.

[0011] The chlorophenol-containing wastewater from upstream enters the activated carbon adsorption tower from the bottom through a pipeline. The outlet of the activated carbon adsorption tower is connected to the downstream purified water network. Fresh activated carbon enters from the top of the activated carbon adsorption tower, while waste activated carbon is discharged from the bottom and transported to the downstream thermal desorption system via conveying water. During this process, excess conveying water is separated by a water tank and then pumped back to the conveying water network by a circulating water pump.

[0012] The thermal desorption system includes inert desorption gas from upstream, a desorption rotary kiln, and regenerated activated carbon;

[0013] Inert desorption gas from upstream enters the desorption rotary kiln through a pipeline. The inert desorption gas is composed of inert gases such as nitrogen and carbon dioxide. The working temperature of the inert desorption gas is 135-600℃, and the working pressure is atmospheric pressure to 1.0 MPaG.

[0014] The catalytic combustion system includes a first oxidant from the utility pipeline network, a second oxidant from the utility pipeline network, a first catalytic burner and a second catalytic burner, and purified gas going downstream;

[0015] Harmful gases from the thermal desorption system enter the first catalytic burner through the outlet pipe of the desorption rotary kiln. The first oxidant from the utility pipeline enters the first catalytic burner through a pipeline. The reacted gas enters the second catalytic burner through a pipeline. The second oxidant from the utility pipeline enters the second catalytic burner through a pipeline. The reacted gas is sent downstream as purified gas.

[0016] The catalytic combustion system operates at atmospheric pressure to 1.0 MPaG, the first catalytic burner operates at a temperature of 150 to 450°C, the second catalytic burner operates at a temperature of 100 to 500°C, and the catalyst used is a dedicated catalyst for dioxin catalytic combustion.

[0017] The intelligent control system includes a wastewater flow control valve, a wastewater flow meter, a wastewater sampler, a purified water flow meter, a purified water sampler, an inert desorption gas flow control valve, an inert desorption gas flow meter, an inert desorption gas thermometer, a desorption rotary furnace outlet thermometer, a first catalytic burner thermometer, a second catalytic burner thermometer, a first oxidant flow control valve, a first oxidant flow meter, a first oxidant oxygen analyzer, a second oxidant flow control valve, a second oxidant flow meter, a second oxidant oxygen analyzer, a purified gas analyzer, a purified gas flow meter, and a purified gas oxygen analyzer; wherein:

[0018] The wastewater flow meter, wastewater sampler, and purified water flow meter, as well as the purified water sampler's measurement values, are all remotely input into the chlorine source calculation module in the central controller.

[0019] The measured values ​​of the inert desorption gas flow meter, the inert desorption gas thermometer, and the desorption rotary kiln outlet thermometer are all remotely input into the desorption calculation module in the central controller.

[0020] The measured values ​​of the first oxidant flow meter, the first oxidant oxygen analyzer, the second oxidant flow meter, the second oxidant oxygen analyzer, the purified gas flow meter, and the purified gas oxygen analyzer are all remotely input into the oxygen source calculation module in the central controller.

[0021] The measured values ​​from the first and second catalytic burner thermometers are remotely input into the catalytic calculation module in the central controller.

[0022] The calculated values ​​from the chlorine source calculation module, desorption calculation module, oxygen source calculation module, and catalysis calculation module are all fed into the data dimensionality reduction module. Their output values ​​are fed into the normalization module, the neural network module, the logic control module, and the valve control module.

[0023] Specifically, the dimensionality reduction methods used by the data dimensionality reduction module include, but are not limited to, principal component analysis (PCA), linear discriminant analysis (LDA), and other commonly used data dimensionality reduction methods in machine learning.

[0024] The neural network module has 1 to 3 input layer units, 2 to 4 hidden layer units, 1 to 4 hidden layers, and 1 to 2 output layer units.

[0025] Furthermore, the measured values ​​from the purified gas analyzer are input into the neural network module as training sample data for the neural network model.

[0026] The control method for the above-mentioned wastewater chlorophenol compound harmless treatment device is characterized by comprising the following steps:

[0027] Step 1: Chlorophenol-containing wastewater from upstream enters an activated carbon adsorption tower, where chlorophenol compounds are enriched in the activated carbon. A wastewater flow meter measures the wastewater flow rate, and a wastewater sampler determines the chlorine content in the wastewater. The purified water after adsorption enters the downstream system, where a purified water flow meter measures the purified water flow rate, and a purified water sampler determines the chlorine content in the purified water. These measurements are remotely input into the chlorine source calculation module in the central controller to calculate the chlorine content.

[0028] Step 2: Waste activated carbon is fed into the desorption rotary kiln via a constant flow of water. Inert desorption gas enters the desorption rotary kiln to heat and desorb the waste activated carbon. The resulting high-concentration harmful gases enter the catalytic combustion system. The flow rate measured by the inert desorption gas flow meter and the temperature measured by the inert desorption gas thermometer are remotely input into the desorption calculation module in the central controller to calculate the desorption heat load.

[0029] Step 3: The harmful gases generated by the thermal desorption system sequentially enter the first catalytic burner, the second catalytic burner, and the oxidant to carry out catalytic combustion reactions and generate different catalyst bed temperatures. The thermometers of the first and second catalytic burners measure the catalyst bed temperature respectively, and the temperature measurement values ​​are remotely input into the catalytic calculation module in the central controller to calculate the catalytic reaction depth.

[0030] Step 4: The depth of the catalytic reaction is controlled by the oxidant dosage. The measured values ​​of the first oxidant flow meter, the first oxidant oxygen analyzer, the second oxidant flow meter, the second oxidant oxygen analyzer, the purified gas flow meter, and the purified gas oxygen analyzer are remotely input into the oxygen source calculation module in the central controller to calculate the oxidant dosage input into the catalytic combustion system.

[0031] Step 5: The amount of chlorine, desorption heat load, catalytic reaction depth and oxidation dosage are fed into the data dimensionality reduction module to reduce the dimensionality of the data, reducing the original 4-dimensional data to 1-3 dimensions, which is used to reduce the training complexity of the neural network module and reduce training error.

[0032] Step 6: The dimensionality-reduced data enters the normalization module to complete the data normalization, converting it into dimensionless values ​​in the range of 0 to 1;

[0033] Step 7: The normalized values ​​are fed into the neural network module for predictive model training. The actual measured values ​​of the purified gas analyzer continuously correct the neural network model, and finally the training and construction of the neural network model are completed.

[0034] Step 8: After the network model is built, the newly measured chlorine amount, desorption heat load, catalytic reaction depth and oxidation dosage are entered into the neural network model according to steps 5 and 6, and its output value enters the logic control module.

[0035] Step 9: The logic control module compares the predicted value with the set value and inputs the compared data into the valve control module. The valve control module outputs the wastewater flow rate set value, the inert desorption gas flow rate set value, the first oxidant flow rate set value, and the second oxidant flow meter set value to control the wastewater flow control valve, the inert desorption gas flow control valve, the first oxidant flow control valve, and the second oxidant flow control valve, respectively, thus completing the automatic regulation of a wastewater chlorophenol compound harmless treatment device.

[0036] The wastewater chlorophenol compound harmless treatment device and control method of the present invention decomposes the generated dioxin-like substances through a dedicated catalytic combustion technology, and completes automatic control and precise adjustment through an intelligent control system. This solves engineering problems such as the difficulty in measuring dioxin concentration and the complexity of dioxin generation mechanisms. Compared with existing technologies, the technological innovation and advantages are as follows:

[0037] 1. An activated carbon regeneration process is adopted. The regenerated activated carbon generated by inert desorption gas is then used for downstream activation treatment. The activated regenerated activated carbon can be mixed with fresh activated carbon to continue adsorbing chlorophenol-containing wastewater. Since the regeneration cost of activated carbon is much lower than the manufacturing cost of fresh activated carbon, this device, as an upstream technology for activated carbon activation, can reduce the operating cost of the wastewater treatment process.

[0038] 2. A process approach of enriching and rendering harmless chlorophenol compounds was adopted. First, activated carbon enriches the low-concentration chlorophenols in the wastewater. Then, the saturated activated carbon undergoes desorption treatment. Since chlorophenols can desorb and volatilize below 300℃, the concentration of chlorophenols in the desorbed gas will be much higher than that in the wastewater, thus completing the enrichment of chlorophenols. Given that the catalytic combustion of chlorophenols requires a certain concentration, the high-concentration chlorophenol-containing desorbed gas produced through activated carbon adsorption followed by desorption meets the concentration requirements for catalytic combustion.

[0039] 3. This solution addresses the issue of dioxins generated during the incomplete combustion of spent activated carbon in coal-fired boilers. Conventional boiler systems often lack dedicated dioxin removal devices, only having basic desulfurization and denitrification infrastructure. Some boilers are equipped with quenching devices intended to rapidly cool dioxin precursors and inhibit their further reaction, but these methods are often ineffective. By using a dedicated catalyst for dioxin catalytic combustion, the dioxins can be decomposed, producing harmless byproducts.

[0040] 4. Automated monitoring and intelligent control of the device were achieved. Due to the complexity of dioxin formation mechanisms and the long concentration detection cycle, and the fact that chloride content in wastewater, input heat of inert desorption gas, oxidant concentration, and catalyst bed temperature all affect dioxin formation, it is a multivariate control system. Since the dioxin formation mechanism model is immature, this invention uses a black-box model such as a neural network. Through data dimensionality reduction, the number of samples required for model training is reduced, ultimately completing the construction of the neural network model and achieving rapid soft measurement of dioxin concentration and rapid device regulation. Attached Figure Description

[0041] Figure 1 This is a flow chart of the wastewater chlorophenol compound harmless treatment device of the present invention.

[0042] Figure 2 This is a schematic diagram of the data link for the harmless treatment and control method of chlorophenol compounds in wastewater according to the present invention.

[0043] Figure 3 This is a schematic diagram of a neural network model.

[0044] The markings in the diagram represent: 1. Activated carbon adsorption tower; 2. Circulating water pump; 3. Water delivery tank; 4. Desorption rotary kiln; 5. First catalytic burner; 6. Second catalytic burner; 7. Wastewater; 8. Purified water; 9. Fresh activated carbon; 10. Delivery water; 11. Inert desorption gas; 12. Regenerated activated carbon; 13. First oxidant; 14. Second oxidant; 15. Purified gas; 16. Wastewater flow control valve; 17. Wastewater flow meter; 18. Wastewater sampler; 19. Purified water flow meter; 20. Purified water sampler; 21. Inert desorption gas flow control valve; 22. Inert desorption gas flow meter; 23. Inert desorption gas thermometer; 24. Desorption rotary kiln outlet thermometer; 25. First oxidant flow control valve; 26. First oxidant flow meter; 27. First oxidant oxygen analyzer; 28. First catalytic burner thermometer. 29. Second oxidant flow control valve; 30. Second oxidant flow meter; 31. Second oxidant oxygen analyzer; 32. Second catalytic burner thermometer; 33. Purified gas analyzer; 34. Purified gas flow meter; 35. Purified gas oxygen analyzer; 36. Chlorine source calculation module; 37. Desorption calculation module; 38. Oxygen source calculation module; 39. Catalysis calculation module; 40. Data dimensionality reduction module; 41. Normalization module; 42. Neural network module; 43. Logic control module; 44. Valve control module; 45. Input layer; 46. Hidden layer; 47. Output layer; 48. Chlorine element quantity; 49. Desorption heat load; 50. Catalytic reaction depth; 51. Oxidation dosage; 52. Wastewater flow setpoint; 53. Inert desorption gas flow setpoint; 54. First oxidant flow setpoint; 55. Second oxidant flow setpoint.

[0045] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. Detailed Implementation

[0046] See Figure 1 , Figure 2 and Figure 3 This embodiment provides a device for the harmless treatment of chlorophenol compounds in wastewater, comprising a wastewater adsorption system, a thermal desorption system, a catalytic combustion system, and an intelligent control system, wherein:

[0047] The wastewater adsorption system includes 7 chlorophenol wastewater from upstream, 1 activated carbon adsorption tower, 8 purified water going downstream, 9 fresh activated carbon from upstream, 10 transported water from the public works pipeline network, 2 circulating water pump and 3 transport water distribution tank.

[0048] Chlorinated phenol wastewater 7 from upstream enters activated carbon adsorption tower 1 from the bottom through a pipeline. The outlet of activated carbon adsorption tower 1 is connected to the downstream purified water network 8. Fresh activated carbon 9 enters from the top of activated carbon adsorption tower 1, while waste activated carbon is discharged from the bottom and transported to the downstream thermal desorption system via conveying water 10. During this process, excess conveying water is separated by conveying water tank 3 and then transported back to the conveying water network 10 by circulating water pump 2.

[0049] The working principle of the wastewater adsorption system is as follows: wastewater and activated carbon layer come into countercurrent contact. The wastewater passes through the activated carbon layer from bottom to top. Chlorophenols in the wastewater are continuously adsorbed by the activated carbon. The treated wastewater leaves from the top of the activated carbon adsorption tower.

[0050] The thermal desorption system includes inert desorption gas 11 from upstream, a desorption rotary kiln 4, and regenerated activated carbon 12;

[0051] Inert desorption gas 11 from upstream enters the desorption rotary kiln 4 through a pipeline. The type of inert desorption gas 11 is nitrogen. The working temperature of inert desorption gas 11 is 600℃ and the working pressure is atmospheric pressure.

[0052] Working principle of thermal desorption system: High temperature inert desorption gas enters the desorption rotary furnace and comes into contact with saturated activated carbon. Due to the heat, the chlorophenols adsorbed in the activated carbon pores are desorbed from the activated carbon pores.

[0053] The catalytic combustion system includes a first oxidant 13 from the utility pipeline network, a second oxidant 14 from the utility pipeline network, a first catalytic burner 5 and a second catalytic burner 6, and purified gas 15 going downstream.

[0054] Working principle of catalytic combustion system: Due to the presence of catalyst, the combustion oxidation temperature of the exhaust gas generated during the desorption process is greatly reduced. In catalytic combustion system, air and desorbed exhaust gas enter the catalyst bed with the temperature controlled within a certain range for oxidation reaction.

[0055] Harmful gases from the thermal desorption system enter the first catalytic burner 5 through the outlet pipe of the desorption rotary kiln. The first oxidant 13 from the utility pipeline enters the first catalytic burner 5 through the pipeline. The gas after reaction enters the second catalytic burner 6 through the pipeline. The second oxidant 14 from the utility pipeline enters the second catalytic burner 6 through the pipeline. The gas after reaction is sent downstream as purified gas 15.

[0056] The catalytic combustion system operates at atmospheric pressure. The first catalytic burner 5 operates at 250°C, and the second catalytic burner 6 operates at 350°C. The catalyst used is a dioxin catalytic combustion-specific catalyst, whose main component is titanium.

[0057] The intelligent control system includes a wastewater flow control valve 16, a wastewater flow meter 17, a wastewater sampler 18, a purified water flow meter 19, a purified water sampler 20, an inert desorption gas flow control valve 21, an inert desorption gas flow meter 22, an inert desorption gas thermometer 23, a desorption rotary kiln outlet thermometer 24, a first catalytic burner thermometer 28, a second catalytic burner thermometer 32, a first oxidant flow control valve 25, a first oxidant flow meter 26, a first oxidant oxygen analyzer 27, a second oxidant flow control valve 29, a second oxidant flow meter 30, a second oxidant oxygen analyzer 31, a purified gas analyzer 33, a purified gas flow meter 34, and a purified gas oxygen analyzer 35.

[0058] In this embodiment, the measured values ​​of the wastewater flow meter 17, wastewater sampler 18, purified water flow meter 19, and purified water sampler 20 are all remotely input into the chlorine source calculation module 36 in the central controller.

[0059] The measured values ​​of the inert desorption gas flow meter 22, the inert desorption gas thermometer 23, and the desorption rotary furnace outlet thermometer 24 are all remotely input into the desorption calculation module 37 in the central controller.

[0060] The measured values ​​of the first oxidant flow meter 26, the first oxidant oxygen analyzer 27, the second oxidant flow meter 30, the second oxidant oxygen analyzer 31, the purified gas flow meter 34, and the purified gas oxygen analyzer 35 are all remotely input into the oxygen source calculation module 38 in the central controller; the measured values ​​of the first catalytic burner thermometer 28 and the second catalytic burner thermometer 32 are all remotely input into the catalytic calculation module 39 in the central controller.

[0061] The calculated values ​​of the four modules, namely chlorine source calculation module 36, desorption calculation module 37, oxygen source calculation module 38 and catalysis calculation module 39, are all entered into the data dimensionality reduction module 40. The output values ​​are entered into the normalization module 41, the neural network module 42, the logic control module 43, and the valve control module 44.

[0062] In this embodiment, the data dimensionality reduction module 40 uses principal component analysis (PCA) as the dimensionality reduction method.

[0063] In neural network module 42, the input layer 45 has 2 units, the hidden layer 46 has 4 units, the hidden layer has 2 units, and the output layer 47 has 1 unit.

[0064] The measured values ​​of the purified gas analyzer 33 are remotely input into the neural network module 42 in the central controller as training parameters for the neural network model.

[0065] The control method for the above-mentioned wastewater chlorophenol compound harmless treatment device includes the following steps:

[0066] Step 1: Chlorophenol-containing wastewater 7 from upstream enters activated carbon adsorption tower 1, where chlorophenol compounds are enriched in the activated carbon. Wastewater flow meter 17 measures the flow rate of wastewater 7, and wastewater sampler 18 determines the chlorine content in wastewater 7. The purified water 8 after adsorption enters the downstream system. Purified water flow meter 19 measures the flow rate of purified water 8, and purified water sampler 20 determines the chlorine content in purified water 8. The measured values ​​are remotely input into the chlorine source calculation module 36 in the central controller to calculate the chlorine content 48.

[0067] Calculation of chlorine element 48: The flow rate measured by wastewater flow meter 17 is F1, the concentration measured by wastewater sampler 18 is C1, the concentration measured by purified water flow meter 19 is F2, the concentration measured by purified water sampler 20 is C2, and the chlorine element 48 is M. Then M = F1 × C1 - F2 × C2.

[0068] Step 2: Waste activated carbon enters the desorption rotary kiln 4 through a constant flow of water 10. Inert desorption gas 11 enters the desorption rotary kiln 4 to heat and desorb the waste activated carbon. The resulting high-concentration harmful gas enters the catalytic combustion system. The flow rate measured by the inert desorption gas flow meter 22, the temperature measured by the inert desorption gas thermometer 23, and the temperature measured by the desorption rotary kiln outlet thermometer 24 are remotely input into the desorption calculation module 37 in the central controller to calculate the desorption heat load 49. The desorption heat load 49 is calculated as follows: the flow rate measured by the inert desorption gas flow meter 22 is F, the temperature measured by the inert desorption gas thermometer 23 is T1, the temperature measured by the desorption rotary kiln outlet thermometer 24 is T2, and the desorption heat load is Q. Then, Q = F × (T2 - T1).

[0069] Step 3: The harmful gases generated by the thermal desorption system sequentially enter the first catalytic burner 5, the second catalytic burner 6, and the oxidant to carry out catalytic combustion reactions and generate different catalyst bed temperatures. The first catalytic burner thermometer 28 and the second catalytic burner thermometer 32 measure the catalyst bed temperature respectively. The temperature measurement values ​​are remotely input into the catalytic calculation module 39 in the central controller to calculate the catalytic reaction depth 50. The catalytic reaction depth 50 is calculated as follows: the temperature measured by the first catalytic burner thermometer 28 is T1, the temperature measured by the second catalytic burner thermometer 32 is T2, and the reaction depth is η. Then η = T2 - T1.

[0070] Step 4: The depth of the catalytic reaction is controlled by the oxidant dose. The measured values ​​of the first oxidant flow meter 26, the first oxidant oxygen analyzer 27, the second oxidant flow meter 30, the second oxidant oxygen analyzer 31, the purified gas flow meter 34, and the purified gas oxygen analyzer 35 are remotely input into the oxygen source calculation module 38 in the central controller to calculate the oxidant dose 51 input to the catalytic combustion system. The oxidant dose 51 is calculated as follows: the measured value of the first oxidant flow meter 26 is F1, the measured value of the first oxidant oxygen analyzer 27 is C1, the measured value of the second oxidant flow meter 30 is F2, the measured value of the second oxidant oxygen analyzer 31 is C2, the measured value of the purified gas flow meter 34 is F3, and the measured value of the purified gas oxygen analyzer 35 is C3. The oxidant dose is O, and O = F1*C1 + F2*C2 - F3*C3.

[0071] Step 5: Chlorine element amount 48, desorption heat load 49, catalytic reaction depth 50, and oxidation dosage 51 are entered into the data dimensionality reduction module 40 to reduce the data dimensionality, reducing the original 4-dimensional data to 2-dimensional.

[0072] Step 6: The dimensionality-reduced data enters the normalization module 41 to complete the normalization of the data, converting it into dimensionless values ​​in the range of 0 to 1;

[0073] Step 7: The normalized values ​​are fed into the neural network module 42 for machine learning. The actual measured values ​​of the purified gas analyzer 33 are normalized and then fed into the neural network module 42 for model correction and training. The neural network model is continuously corrected and finally completed. The purified gas analyzer 33 is equipped with a gas chromatograph-mass spectrometer as needed, which can detect the dioxin content in flue gas. The neural network suitable for this invention is a backpropagation neural network.

[0074] Step 8: The network model is completed (the criteria for judging the completion of the network model is: the error between the normalized predicted value of the model and the normalized value of the actual measurement is <5%). The newly measured chlorine element amount 48, desorption heat load 49, catalytic reaction depth 50 and oxidation dosage 51 are entered into the neural network model after Step 5 and Step 6, and their output values ​​are entered into the logic control module 43.

[0075] Step 9: The logic control module 43 compares the predicted value and the measured value, and inputs the compared data into the valve control module. The valve control module outputs the wastewater flow rate setpoint 52, the inert desorption gas flow rate setpoint 53, the first oxidant flow rate setpoint 54, and the second oxidant flow meter setpoint 55 to control the wastewater flow rate control valve 16, the inert desorption gas flow rate control valve 21, the first oxidant flow rate control valve 25, and the second oxidant flow rate control valve 29, respectively, and finally completes the automatic regulation of a wastewater chlorophenol compound harmless treatment device.

[0076] Application Examples:

[0077] The wastewater chlorophenol compound harmless treatment device of this embodiment is used to control a certain chlorophenol-containing wastewater harmless treatment system. The neural network model used is a backpropagation neural network. The purified gas analyzer 33 selects an isotope dilution high-resolution gas chromatography-high-resolution mass spectrometry to detect the dioxin content in the flue gas offline.

[0078] The process and completion conditions for network training in this embodiment are as follows: when the deviation σ < 5%, the neural network model training is completed.

[0079] Step 1: Model Training

[0080] (1) Data Acquisition: The data acquisition time was 1 hour, that is, the average values ​​of four sets of process parameters were continuously measured and calculated: chlorine content 48 was 125 ppm, desorption heat load 49 was 1.5 MJ / h, catalytic reaction depth 50 was 0.73, and oxidation dosage 51 was 21%; the actual dioxin data collected by the purified gas analyzer 33 was 0.1 TEQ / m³. 3 As training sample 1; with chlorine element content of 48 at 130 ppm, desorption heat load of 49 at 1.5 MJ / h, catalytic reaction depth of 50 at 0.73, and oxidation dose of 51 at 21%; the actual data collected by the purified gas analyzer (33) was 0.12 TEQ / m 3 As training sample 2; a total of 20 sets of data were collected as training samples, and four data points other than dioxin content were used as the initial input data in this embodiment.

[0081] The following description of the neural network training process uses only Sample 1 as an example.

[0082] (2) Data dimensionality reduction: Principal component analysis was used to reduce the four-element single-row matrix [1251.50.7321] of sample 1 to a two-element single-row matrix, that is, the number of principal components is 2. The data matrix after dimensionality reduction is:

[25317] . The dimensionality reduction result is related to the order of the mean values ​​of the above four groups of data. However, as long as the order is fixed, it can be arranged in the order of 48 / 49 / 50 / 51. This order can be followed every time. (3) Normalization: Normalization is performed using the largest number in the matrix as the benchmark value to obtain: [0.081]. This data is input into the neural network model for training and the calculated value is output. In this embodiment, the initial model of the neural network model is ax1+bx2, where a and b are weights. The initial weight values ​​are set as follows: a=1, b=0.01. The calculated dioxin content is 1×0.08+0.01×1=0.09.

[0083] (4) Deviation calculation: At this time, the actual data collected by the purified gas analyzer 33 is: 0.1 TEQ / m³ 3 The calculated deviation value is σ = (0.1 - 0.09) / 0.1 * 100% = 10%. The training is completed when the deviation σ < 5%. Since 10% > 5%, model learning needs to continue.

[0084] (5) Weight adjustment: Since the above calculation found that the deviation σ > 5%, the weights of the neural network model are adjusted to a = 1.1 and b = 0.01;

[0085] (6) Using 20 sets of data in this embodiment as training samples, after training the neural network 1000 times, the model weights are finally adjusted to a = 1.11, b = 0.01, σ < 5%. After training, the final neural network model is 1.11x1 + 0.01x2.

[0086] Step 2: Neural Network Prediction of Dioxin Content

[0087] (1) Data acquisition: At this time, the average real-time data for 1 hour is chlorine element amount 48 is 145 ppm, desorption heat load 49 is 1.5 MJ / h, catalytic reaction depth 50 is 0.73, and oxidation dose (51) is 21%.

[0088] (2) Data dimensionality reduction: Principal component analysis was used to reduce the four-element single-row matrix [1451.50.7321] of sample 1 to a two-element single-row matrix, that is, the number of principal components is 2. The data matrix after dimensionality reduction is:

[31315] .

[0089] (3) Normalization: Normalize using the largest number in the matrix as the base value to get: [0.0981].

[0090] (4) Calculate the neural network prediction value of dioxin content: 1.11 × 0.098 + 0.01 × 1 = 0.119 TEQ / m 3 .

[0091] Step 3: Logic control of the valve

[0092] (1) Deviation calculation: The emission setpoint for dioxins is set at 0.1 TEQ / m³. 3 The pre-test result in the above embodiment is 0.119 TEQ / m 3 If the deviation is higher than the set value, the deviation Σ=(0.119-0.1) / 0.1×100%=19%, and this deviation is input to the valve control module 44;

[0093] (2) Valve opening adjustment: As shown in the previous step, the dioxin content has increased significantly. The dioxin emission needs to be reduced by the following methods: reduce the opening of the upstream wastewater flow control valve 16, increase the opening of the inert desorption gas flow control valve 21, increase the opening of the first oxidant flow control valve 25, and increase the opening of the second oxidant flow control valve 29 to reduce the dioxin emission to the set value.

[0094] According to the above embodiment, the wastewater flow rate setpoint 52 is adjusted to 125m³. 3 / h; the inert desorption gas flow rate setpoint 53 is adjusted to 14500 Nm. 3 / h; The first oxidant flow rate setpoint 54 is adjusted to 25m. 3 / h; The second oxidant flow rate setpoint 55 is adjusted to 27m. 3 / h. Correspondingly, the opening degree of the upstream wastewater flow control valve 16 is reduced by 5%, the opening degree of the inert desorption gas flow control valve 21 is increased by 10%, the opening degree of the first oxidant flow control valve 25 is increased by 2%, and the opening degree of the second oxidant flow control valve 29 is increased by 1%.

[0095] After applying the control concept of this application, compared to before its use, the device's performance is as follows: Chlorophenol wastewater is commonly treated industrially using activated carbon adsorption. However, the adsorbed activated carbon is enriched with large amounts of chlorophenols, typically exceeding 10,000 ppm. Therefore, a large amount of dioxins is generated during the subsequent thermal regeneration of the activated carbon. In traditional flue gas treatment schemes, dioxins are decomposed at high temperatures in the secondary combustion chamber and then rapidly cooled to below 200°C to prevent their resynthesis. Subsequent removal is achieved through secondary adsorption of activated carbon powder. However, this approach is ineffective, and the secondary removal by activated carbon powder still represents a transfer of harmful substances rather than complete treatment. Therefore, in this embodiment, although the hot flue gas undergoes catalytic combustion under the action of a dioxin-specific catalyst, this catalytic combustion process is influenced by multiple factors, including the upstream flue gas temperature, the dioxin content in the flue gas, and the oxygen concentration during catalytic combustion. Since the ultimate goal of this application is to achieve the harmless disposal of a chlorophenol compound, the key indicator is the measurement of dioxin content. The traditional method is offline measurement, which involves sampling the flue gas for several hours and then testing the gas sample through an analytical instrument. This process usually takes several hours to complete. Therefore, the detection of dioxins has never been able to achieve the goal of rapid, economical, accurate, and online measurement.

[0096] To address the aforementioned issues, the inventors initially used soft sensing methods to calculate dioxin content during their research. However, dioxin content is the result of controlling multiple process variables, including chlorine content at the process source, heat of thermal regeneration, reaction progress, and oxygen content. Traditional control methods cannot effectively meet the demands of multi-variable control. Therefore, the inventors ultimately used an artificial neural network model—a black box model—which effectively solved the aforementioned problems.

[0097] In this embodiment, by using 10-20 sets of real dioxin analysis and detection values ​​as training parameters to input into the artificial neural network model, a correspondence is gradually established between the four input variables and the dioxin measurement values, ultimately enabling the model to perfectly predict the dioxin content.

[0098] Finally, the wastewater flow rate setpoint, inert desorption gas flow rate setpoint, first oxidant flow rate setpoint, and second oxidant flow rate setpoint are calculated using the dioxin concentration output by the artificial neural network. These setpoints are then input into the actuators of the relevant process valves to ultimately meet the process parameter conditions, ensuring that the wastewater chlorophenol compound harmless treatment device in this embodiment operates automatically and stably, achieving complete dioxin treatment.

Claims

1. A device for the harmless treatment of chlorophenol compounds in wastewater, characterized in that, It includes a wastewater adsorption system, a thermal desorption system, a catalytic combustion system, and an intelligent control system, among which: The wastewater adsorption system includes chlorophenol wastewater (7) from upstream, activated carbon adsorption tower (1), purified water going downstream (8), fresh activated carbon (9) from upstream, water transported from public utility network (10), circulating water pump (2) and water transport water distribution tank (3). Chlorinated phenol wastewater (7) from upstream enters the activated carbon adsorption tower (1) from the bottom through a pipeline. The outlet of the activated carbon adsorption tower (1) is connected to the downstream purified water (8) network. Fresh activated carbon (9) enters from the top of the activated carbon adsorption tower (1), and waste activated carbon is discharged from the bottom and transported to the downstream thermal desorption system through the transport water (10). During this period, excess transport water is separated by the transport water tank (3) and transported back to the transport water (10) network by the circulating water pump (2). The thermal desorption system includes inert desorption gas (11) from upstream, a desorption rotary kiln (4) and regenerated activated carbon (12); Inert desorption gas (11) from upstream enters the desorption rotary kiln (4) through a pipeline. The inert desorption gas (11) is nitrogen, carbon dioxide and other inert gases. The working temperature of the inert desorption gas (11) is 135℃~600℃ and the working pressure is atmospheric pressure~1.0MPaG. The catalytic combustion system includes a first oxidant (13) from the utility pipeline network, a second oxidant (14) from the utility pipeline network, a first catalytic burner (5) and a second catalytic burner (6) and purified gas (15) going downstream; Harmful gases from the thermal desorption system enter the first catalytic burner (5) through the outlet pipe of the desorption rotary kiln. The first oxidant (13) from the utility pipeline enters the first catalytic burner (5) through the pipeline. The gas after reaction enters the second catalytic burner (6) through the pipeline. The second oxidant (14) from the utility pipeline enters the second catalytic burner (6) through the pipeline. The gas after reaction is sent downstream as purified gas (15). The working pressure of the catalytic combustion system is atmospheric pressure to 1.0 MPaG, the working temperature of the first catalytic burner (5) is 150℃ to 450℃, the working temperature of the second catalytic burner (6) is 100℃ to 500℃, and the catalyst used is a dioxin catalytic combustion specific catalyst. The intelligent control system includes a wastewater flow control valve (16), a wastewater flow meter (17), a wastewater sampler (18), a purified water flow meter (19), a purified water sampler (20), an inert desorption gas flow control valve (21), an inert desorption gas flow meter (22), an inert desorption gas thermometer (23), a desorption rotary kiln outlet thermometer (24), a first catalytic burner thermometer (28), a second catalytic burner thermometer (32), a first oxidant flow control valve (25), a first oxidant flow meter (26), a first oxidant oxygen analyzer (27), a second oxidant flow control valve (29), a second oxidant flow meter (30), a second oxidant oxygen analyzer (31), a purified gas analyzer (33), a purified gas flow meter (34), and a purified gas oxygen analyzer (35).

2. The wastewater chlorophenol compound harmless treatment device as described in claim 1, characterized in that, The measured values ​​of the wastewater flow meter (17), wastewater sampler (18), purified water flow meter (19), and purified water sampler (20) are all remotely input into the chlorine source calculation module (36) in the central controller. The measured values ​​of the inert desorption gas flow meter (22), the inert desorption gas thermometer (23), and the desorption rotary furnace outlet thermometer (24) are all remotely input into the desorption calculation module (37) in the central controller; The measured values ​​of the first oxidant flow meter (26), the first oxidant oxygen analyzer (27), the second oxidant flow meter (30), the second oxidant oxygen analyzer (31), the purified gas flow meter (34), and the purified gas oxygen analyzer (35) are all remotely input into the oxygen source calculation module (38) in the central controller. The measured values ​​of the first catalytic burner thermometer (28) and the second catalytic burner thermometer (32) are remotely input into the catalytic calculation module (39) in the central controller; The calculated values ​​of the four modules, namely the chlorine source calculation module (36), the desorption calculation module (37), the oxygen source calculation module (38), and the catalysis calculation module (39), are all entered into the data dimensionality reduction module (40), its output value enters the normalization module (41), its output value enters the neural network module (42), its output value enters the logic control module (43), and its output value enters the valve control module (44).

3. The wastewater chlorophenol compound harmless treatment device as described in claim 2, characterized in that, The dimensionality reduction methods used in the data dimensionality reduction module (40) include, but are not limited to, principal component analysis (PCA) and linear discriminant analysis (LDA), which are commonly used data dimensionality reduction methods in machine learning.

4. The wastewater chlorophenol compound harmless treatment device as described in claim 2, characterized in that, In the neural network module (42), the number of units in the input layer (45) is 1 to 3, the number of units in the hidden layer (46) is 2 to 4, the number of units in the hidden layer is 1 to 4, and the number of units in the output layer (47) is 1 to 2.

5. The wastewater chlorophenol compound harmless treatment device as described in claim 1, characterized in that, The measured values ​​of the purified gas analyzer (33) are input into the neural network module (42) as training parameters for the neural network model.

6. The control method of the wastewater chlorophenol compound harmless treatment device according to any one of claims 1 to 5, characterized in that, Includes the following steps: Step 1: Chlorophenol-containing wastewater (7) from upstream enters activated carbon adsorption tower (1), where chlorophenol compounds are enriched in activated carbon. Wastewater flow meter (17) is used to measure the flow rate of wastewater (7), and wastewater sampler (18) is used to determine the chlorine content in wastewater (7). After adsorption, purified water (8) enters the downstream system. Purified water flow meter (19) is used to measure the flow rate of purified water (8), and purified water sampler (20) is used to determine the chlorine content in purified water (8). The above measured values ​​are remotely input into the chlorine source calculation module (36) in the central controller to calculate the amount of chlorine (48). Step 2: Waste activated carbon enters the desorption rotary furnace (4) through a constant flow of water (10), and inert desorption gas (11) enters the desorption rotary furnace (4) to heat and desorb the waste activated carbon. The high concentration of harmful gas generated enters the catalytic combustion system. The flow rate measured by the inert desorption gas flow meter (22) and the temperature measured by the inert desorption gas thermometer (23) are remotely input into the desorption calculation module (37) in the central controller to calculate the desorption heat load (49). Step 3: The harmful gases generated by the thermal desorption system enter the first catalytic burner (5), the second catalytic burner (6) and the oxidant in sequence to carry out catalytic combustion reaction and generate different catalyst bed temperatures. The first catalytic burner thermometer (28) and the second catalytic burner thermometer (32) measure the catalyst bed temperature respectively. The temperature measurement value is remotely input into the catalytic calculation module (39) in the central controller to calculate the catalytic reaction depth (50). Step 4: The depth of the catalytic reaction is controlled by the oxidant dosage. The measured values ​​of the first oxidant flow meter (26), the first oxidant oxygen analyzer (27), the second oxidant flow meter (30), the second oxidant oxygen analyzer (31), the purified gas flow meter (34), and the purified gas oxygen analyzer (35) are remotely input into the oxygen source calculation module (38) in the central controller to calculate the oxidant dosage (51) input into the catalytic combustion system. Step 5: The amount of chlorine (48), desorption heat load (49), catalytic reaction depth (50) and oxidation dosage (51) are entered into the data dimensionality reduction module (40) to reduce the dimensionality of the data, reducing the original 4-dimensional data to 1-3 dimensions, which is used to reduce the training complexity of the neural network module (42) and reduce the training error. Step 6: The dimension-reduced data enters the normalization module (41) to complete the normalization of the data and convert it into a dimensionless value in the range of 0 to 1; Step 7: The normalized values ​​are fed into the neural network module (42) for prediction model training. The actual measured values ​​of the purified gas analyzer (33) are used as training samples to continuously correct the neural network model and finally complete the training and construction of the neural network model. Step 8: After the network model is built, the real-time measured chlorine amount (48), desorption heat load (49), catalytic reaction depth (50) and oxidation dose (51) are entered into the neural network model after Step 5 and Step 6, and their output values ​​are entered into the logic control module (43). Step 9: The logic control module (43) compares the predicted value and the set value, and inputs the compared data into the valve control module (44). The valve control module (44) outputs the wastewater flow set value (52), the inert desorption gas flow set value (53), the first oxidant flow set value (54), and the second oxidant flow meter set value (55) to control the wastewater flow control valve (16), the inert desorption gas flow control valve (21), the first oxidant flow control valve (25), and the second oxidant flow control valve (29) respectively, and finally completes the automatic adjustment of a wastewater chlorophenol compound harmless treatment device.

Images