Optimization control method for air evaporation and concentration technology of desulfurization waste water of thermal power plant

Abstract

The invention discloses an optimal control method for a thermal power plant desulfurization wastewater air evaporation concentration technology, which comprises the following steps: S1, multi-source data real-time acquisition: deploying high-precision sensor groups in a concentration tower and at each inlet and outlet, and acquiring air parameters and wastewater parameters in real time; s2, data preprocessing: preprocessing the collected multi-source data to construct a data set; s3, constructing a fuzzy neural network model: constructing a two-layer fuzzy neural network model comprising a first-layer network and a second-layer network based on a fuzzy neural network algorithm; the first-layer network comprises an air parameter treatment sub-network and a wastewater parameter treatment sub-network which are arranged in parallel; the second-layer network is a decision-making layer and outputs and generates the rotating speed of an air fan, the rotating speed of a circulating pump and the concentration ratio; s4, feedback control and real-time adjustment, wherein the output value serves as a control parameter set value, and the air volume, the circulating wastewater flow, the circulating wastewater temperature and the concentration ratio are adjusted. According to the invention, the evaporation efficiency can be obviously improved, the energy consumption is reduced, and scaling is inhibited.

Description

Technical Field

[0001] This invention relates to the field of desulfurization wastewater treatment technology in thermal power plants, specifically to an optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants. Background Technology

[0002] The wastewater generated by wet desulfurization processes in thermal power plants has a complex composition and contains high concentrations of... , Containing heavy metal ions, direct discharge would severely pollute the environment. Air evaporation concentration technology, due to its advantages such as requiring no chemical reagents and its strong adaptability, has gradually become one of the mainstream processes for concentrating and reducing desulfurization wastewater. However, existing technologies have the following problems: Parameter fixation: Traditional methods use fixed control parameters (such as constant air intake and spray volume), which cannot adapt to fluctuations in wastewater composition and environmental changes, resulting in low evaporation efficiency and high energy consumption.

[0003] Scaling and clogging: Due to poor parameter matching, CaSO4 and CaCO3 scale are easily formed in the concentration tower, requiring frequent shutdowns for cleaning, which increases maintenance costs by 30-50%.

[0004] Lack of intelligence: Existing control systems rely on manual experience for adjustments, resulting in slow response and inability to achieve dynamic optimization.

[0005] Existing improved technologies include PID-based temperature control methods, but these do not consider the effects of multi-parameter coupling. Therefore, there is an urgent need for an intelligent control method that can optimize control parameters in real time and adapt to complex operating conditions. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide an optimized control method for the air evaporation and concentration technology of desulfurization wastewater in thermal power plants, which can realize dynamic optimization of key control parameters, significantly improve evaporation efficiency, reduce energy consumption and inhibit scaling.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows.

[0008] An optimized control method for air evaporation and concentration technology of desulfurization wastewater from thermal power plants includes the following steps: S1. Real-time acquisition of multi-source data: High-precision sensor groups are deployed inside the concentration tower and at each inlet and outlet location to collect air parameters and wastewater parameters in real time; S2. Data Preprocessing: Preprocess the collected multi-source data to construct a dataset; S3. Fuzzy Neural Network Model Construction: Based on the fuzzy neural network algorithm, a two-layer fuzzy neural network model is constructed, including a first layer network and a second layer network. The first layer network includes a parallel air parameter processing sub-network and a wastewater parameter processing sub-network, which output air treatment factors and wastewater characteristic factors, respectively. The second layer network is the decision layer, which generates air fan speed, circulation pump speed and concentration ratio based on the input air treatment factors and wastewater characteristic factors. S4. Feedback Control and Real-time Adjustment: The output value of the fuzzy neural network model is used as the control parameter setpoint. Closed-loop control is achieved through the PID controller to control the frequency conversion of the air fan, wastewater circulation pump and heater hot liquid circulation pump, as well as the running time of the desulfurization wastewater make-up pump and concentrate discharge pump, so as to adjust the air volume, circulating wastewater flow rate, circulating wastewater temperature and concentration ratio.

[0009] Preferably, the air parameters collected in step S1 include air temperature, air humidity, and air flow rate. The air temperature includes the temperature at the top, middle, and bottom of the concentration tower, as well as the air inlet and outlet temperatures. The air humidity includes the humidity at the upper middle, middle, and lower middle of the concentration tower, as well as the air inlet and outlet humidity. The air flow rate includes the flow rates at the air inlet and outlet of the concentration tower. The wastewater parameters include the temperature and flow rate of the circulating wastewater, the pH value of the desulfurization wastewater and its concentrate, and the conductivity of the desulfurization wastewater and its concentrate.

[0010] Preferably, in step S1, the air temperature is measured using an infrared temperature sensor with a measurement range of 0-150℃ and an accuracy of ±1℃; the air humidity is measured using a capacitive humidity sensor with a measurement range of 10-90%RH and an accuracy of ±1.5%; the air flow rate is measured using an ultrasonic air flow meter with a measurement range of 0.5-20m / s and an accuracy of ±0.5%; the circulating wastewater temperature is measured using a PT100 platinum resistance temperature sensor with an accuracy of ±1℃; the circulating wastewater flow rate is measured using an electromagnetic flow meter with a measurement range of 0.1-15m / s and an accuracy of ±0.5%; the pH value of the desulfurization wastewater and its concentrate is measured using an online pH meter with a measurement range of 0-14 and an accuracy of ±0.1pH; and the conductivity of the desulfurization wastewater and its concentrate is measured using an online conductivity meter with a measurement range of 0-300mS / cm and an accuracy of ±1%.

[0011] Preferably, the specific process of step S2 is as follows: S21. Data cleaning: Identify and remove missing values ​​and duplicate records; S22. Handling outlier data based on the 3σ principle: If a parameter exceeds the mean ± 3 times the standard deviation for 3 consecutive times, it is marked as an outlier and replaced with the mean of the previous period. S23. Correlation Analysis and Principal Component Analysis: Calculate the Pearson correlation coefficient matrix to quantify the correlation between parameters; perform principal component analysis on highly collinear features, and retain more than 95% of the variance contribution rate after dimensionality reduction; S24. Data normalization processing: Z-score standardization is used to eliminate dimensional differences. The formula is: in, The output value is the standardized value; σ represents the single original value to be standardized; μ represents the mean of the parameter dataset; σ represents the standard deviation of the parameter dataset. S25. Construct the dataset.

[0012] Preferably, the air parameter processing subnetwork in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 3 nodes, each node's input variables correspond one-to-one with air humidity, air temperature and air flow; The fuzzing layer has 9 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 27 nodes, implementing 3×3×3 fuzzy rules; Normalization layer: 27 nodes, used to normalize the activation degree of rules; Deblurring layer: 1 node, which is the air processing factor. It uses the centroid method to convert the fuzzy output into an accurate value.

[0013] Preferably, the wastewater parameter treatment subnetwork in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 4 nodes, each node has one-to-one input variables: circulating wastewater temperature, circulating wastewater flow rate, pH value of desulfurization wastewater and its concentrate, and conductivity of desulfurization wastewater and its concentrate; The fuzzing layer has 12 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 81 nodes, implementing 3×3×3×3 fuzzy rules; Normalization layer: 81 nodes, used to normalize the activation degree of rules; Defuzzification layer: 1 node, representing wastewater characteristic factors, using the centroid method to convert fuzzy outputs into precise values.

[0014] Preferably, the second layer network in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 2 nodes, each node's input variables correspond one-to-one with air treatment factor and wastewater characteristic factor; The fuzzing layer has 6 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 9 nodes, implementing 3×3 fuzzy rules; Normalization layer: 9 nodes, which normalizes the activation degree of rules; Defuzzification layer: 3 nodes, using the centroid method to convert fuzzy outputs into precise values, with each node's output variable corresponding to the air fan speed, circulating pump speed, and concentration ratio; the concentration ratio is the ratio of the conductivity of the concentrate to the conductivity of the desulfurization wastewater.

[0015] Preferably, the PID closed-loop control logic in step S4 includes: Airflow regulation: Real-time calculation of the deviation between the target airflow and the actual airflow; the PID controller maps different airflows to different frequencies according to the airflow characteristic curve of the air fan, and calculates the air fan frequency adjustment amount based on the deviation value using the PID algorithm; the air fan frequency is dynamically adjusted according to the PID output adjustment amount, so that the actual airflow gradually approaches the target value. Circulating wastewater flow regulation: Based on the flow characteristic curve of the wastewater circulating pump, the target flow rate is mapped to the frequency of the variable frequency, and the frequency of the wastewater circulating pump is dynamically corrected through the PID algorithm to achieve the regulation of the circulating wastewater flow rate. Circulating wastewater temperature regulation: Record the temperature values ​​of the heater hot liquid circulation pump at different frequencies to form a reference for dynamic regulation. Map this to a table showing the correspondence between the target temperature and the frequency of the heater hot liquid circulation pump. The circulating wastewater temperature is regulated by controlling the heat exchange flow rate through the frequency of the heater hot liquid circulation pump. Concentration ratio adjustment: Based on the actual data obtained during commissioning, the concentration ratio is converted into start and stop commands for the desulfurization wastewater make-up pump and the concentrate discharge pump. The concentration ratio is adjusted by controlling the running time of the desulfurization wastewater make-up pump and the concentrate discharge pump.

[0016] Due to the adoption of the above technical solutions, the technical progress achieved by this invention is as follows.

[0017] This invention enables multi-parameter collaborative optimization: traditional methods only focus on a single parameter (such as temperature), while this method integrates temperature, humidity, flow rate, and water quality data to establish a multi-dimensional optimization model, which can greatly improve evaporation efficiency.

[0018] This invention enables intelligent scale prevention: by monitoring the pH, conductivity and temperature of wastewater in real time, the concentration ratio is dynamically adjusted to keep the wastewater saturation below the critical value, thus significantly reducing the scaling rate.

[0019] This invention features fast response: based on FNN and PID closed-loop control, the parameter adjustment response time is ≤5 seconds, greatly improving system stability.

[0020] This method can reduce equipment operating costs and extend equipment life, bringing significant economic benefits to power plants. Attached Figure Description

[0021] Figure 1 This is a system architecture diagram of the present invention.

[0022] Figure 2 This is a flowchart of the data preprocessing process of the present invention.

[0023] Figure 3 This is a diagram of the two-layer fuzzy neural network structure of the present invention.

[0024] Figure 4 This is a diagram of the air parameter processing subnetwork structure of the present invention.

[0025] Figure 5 This is a network structure diagram of the wastewater parameter treatment sub-network of the present invention.

[0026] Figure 6 This is a diagram of the second-layer network structure of the present invention. Detailed Implementation

[0027] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0028] An optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants, combined with Figure 1 As shown, it includes the following steps: S1. Real-time acquisition of multi-source data.

[0029] High-precision sensor arrays are deployed inside the concentration tower and at key locations at each inlet and outlet to collect air and wastewater parameters in real time. Air temperature: An infrared temperature sensor (measurement range 0-150℃, accuracy ±1℃) is used to set multiple air temperature monitoring points at the top, middle, bottom and air inlet and outlet of the concentration tower to obtain air temperature data in real time.

[0030] Air humidity: Using capacitive humidity sensors (measurement range 10-90%RH, accuracy ±1.5%), multiple air humidity monitoring points are set in the upper, middle, lower middle and air inlet and outlet of the concentration tower to obtain the relative humidity inside the tower in real time.

[0031] Air flow rate: Measured at the inlet and outlet of the concentration tower using an ultrasonic air flow meter (measurement range 0.1-20 m / s, accuracy ±0.5%).

[0032] Circulating wastewater temperature and flow rate: A PT100 platinum resistance temperature sensor (accuracy ±1℃) and an electromagnetic flow meter (measurement range 0.1-15 m / s, accuracy ±0.5%) are used to monitor the temperature and flow rate of the wastewater at the inlet and outlet of the concentration tower in real time.

[0033] Wastewater quality: The pH value of desulfurization wastewater and its concentrate is monitored in real time by an online pH meter (measurement range 0-14, accuracy ±0.1pH) to provide early warning of scaling risk; the conductivity of desulfurization wastewater and its concentrate is monitored in real time by an online conductivity meter (measurement range 0-300mS / cm, accuracy ±2%) to reflect the concentration ratio of wastewater in the concentration tower in real time.

[0034] Specifically, air parameters include air temperature, air humidity, and air flow rate; wastewater parameters include circulating wastewater temperature and flow rate, pH value of desulfurization wastewater and its concentrate, and conductivity of desulfurization wastewater and its concentrate. S2. Data preprocessing.

[0035] The collected multi-source data is preprocessed to construct a dataset. The specific process for this step is as follows: S21. Data cleaning: Identify and remove missing values ​​and duplicate records; S22. Handling outlier data based on the 3σ principle: If a parameter exceeds the mean ± 3 times the standard deviation for 3 consecutive times, it is marked as an outlier and replaced with the mean of the previous period. S23. Correlation Analysis and Principal Component Analysis: Calculate the Pearson correlation coefficient matrix to quantify the correlation between parameters. Perform principal component analysis (PCA) on highly collinear features (such as temperature and airflow rate, with a Pearson correlation coefficient r=0.9), and retain more than 95% of the variance contribution rate after dimensionality reduction.

[0036] S24. Data normalization processing: Z-score standardization is used to eliminate dimensional differences. The formula is: in, The output value is the standardized value; σ represents the single original value to be standardized; μ represents the mean of the parameter dataset; and σ represents the standard deviation of the parameter dataset.

[0037] S25. Build a dataset that can be used for model training, testing, and validation.

[0038] S3. Construction of fuzzy neural network model.

[0039] Based on the Fuzzy Neural Network (FNN) algorithm, a mapping model between evaporation efficiency and key control parameters is constructed. Specifically, a two-layer fuzzy neural network model is built, and its two-layer fuzzy neural network structure is as follows: Figure 3 As shown, groups are formed based on parameter relevance and function: The first layer network contains two parallel sub-networks: an air parameter processing sub-network and a wastewater parameter processing sub-network. Details are as follows: Air parameter processing subnetwork: Performs air parameter processing, with inputs of air humidity, air temperature, and air flow rate; output of air processing factor. Its main function is to evaluate the impact of current air conditions on evaporation efficiency.

[0040] Wastewater parameter treatment subnetwork: Performs wastewater parameter treatment. Inputs: circulating wastewater temperature, circulating wastewater flow rate, pH value of desulfurization wastewater and its concentrate, and conductivity of desulfurization wastewater and its concentrate; output: wastewater characteristic factors. Its main function is to evaluate the impact of wastewater characteristics on the evaporation process.

[0041] The second layer network is the decision layer. Inputs include air treatment factors and wastewater characteristic factors. Outputs include air fan speed, circulating pump speed, and concentration ratio. Its main function is to comprehensively consider air and wastewater conditions and output the optimal control parameters.

[0042] Specifically, the air parameter processing subnetwork is a five-layer fuzzy neural network, such as... Figure 4 As shown, the structure is as follows: Input layer: 3 nodes (air humidity, air temperature, and air flow); Fuzzing layer: 9 nodes (3 Gaussian membership functions for each input variable); Rule layer: 27 nodes (3 3 =27 fuzzy rules); Normalization layer: 27 nodes (normalizes the activation of rules); Deblurring layer: 1 node (air processing factor).

[0043] The input layer and fuzzing layer of the air parameter processing subnetwork are as follows: Input layer: Receives three normalized input variables: air humidity (range: 0-1, 0 represents dry, 1 represents humid); air temperature (range: 0-1, 0 represents low temperature, 1 represents high temperature); air flow rate (range: 0-1, 0 represents low flow rate, 1 represents high flow rate).

[0044] Blurring layer: The input is blurred using Gaussian membership functions, and each input variable has 3 Gaussian membership functions (low, medium, and high). The parameters of the Gaussian membership functions (center and width) can be optimized through training.

[0045] The rule layer and normalization layer of the air parameter processing subnetwork: Rule layer: Implements 27 fuzzy rules (3×3×3), as shown in Table 1 below: Table 1: Fuzzy Rules for Air Parameter Processing Serial Number air temperature air humidity airflow Air processing factor Serial Number air temperature air humidity airflow Air processing factor 1 Low Low Low Low 15 middle middle high high 2 Low Low middle Low 16 middle high Low middle 3 Low Low high middle 17 middle high middle high 4 Low middle Low middle 18 middle high high high 5 Low middle middle middle 19 high Low Low Low 6 Low middle high high 20 high Low middle Low 7 Low high Low middle 21 high Low high Low 8 Low high middle high 22 high middle Low Low 9 Low high high high 23 high middle middle middle 10 middle Low Low Low 24 high middle high middle 11 middle Low middle Low 25 high high Low middle 12 middle Low high middle 26 high high middle middle 13 middle middle Low middle 27 high high high middle 14 middle middle middle middle Normalization layer: Normalizes the activation degree of the rules.

[0046] Deblurring layer of the air parameter processing subnetwork: The centroid method is used to convert fuzzy outputs into precise values.

[0047] Specifically, the wastewater parameter treatment subnetwork is a five-layer fuzzy neural network, such as... Figure 5 As shown, the structure is as follows: Input layer: 4 nodes (wastewater temperature, wastewater flow rate, wastewater pH value, and conductivity); fuzzing layer: 12 nodes (3 Gaussian membership functions for each input variable); Rule layer: 81 nodes (3 4 =81 fuzzy rules); Normalization layer: 81 nodes (normalizes the activation degree of rules); Defuzzification layer: 1 node (wastewater characteristic factor).

[0048] The wastewater parameter treatment subnetwork includes its input layer and fuzzification layer. Input layer: Receives four normalized input variables: circulating wastewater temperature (range: 0-1, 0 represents low temperature, 1 represents high temperature), circulating wastewater flow rate (range: 0-1, 0 represents low flow rate, 1 represents high flow rate), pH value of desulfurization wastewater and its concentrate (range: 0-1, 0 represents acidic, 0.5 represents neutral, 1 represents alkaline), and conductivity of desulfurization wastewater and its concentrate.

[0049] Blurring layer: The input is blurred using Gaussian membership functions, and each input variable has 3 Gaussian membership functions (low, medium, and high). The parameters of the Gaussian membership functions (center and width) can be optimized through training.

[0050] The rule layer and normalization layer of the wastewater parameter treatment subnetwork: Rule layer: Implements 81 fuzzy rules (3×3×3×3), as shown in Table 2 below: Table 2: Fuzzy Rules for Wastewater Parameter Treatment Serial Number Circulating wastewater temperature Circulating wastewater flow rate pH value electrical conductivity Wastewater characteristic factors Serial Number Circulating wastewater temperature Circulating wastewater flow rate pH value electrical conductivity Wastewater characteristic factors 1 Low Low acidic Low Low 42 middle middle neutral high Low 2 Low Low acidic middle Low 43 middle middle alkaline Low middle 3 Low Low acidic high Very low 44 middle middle alkaline middle middle 4 Low Low neutral Low Low 45 middle middle alkaline high Low 5 Low Low neutral middle Low 46 middle high acidic Low middle 6 Low Low neutral high Very low 47 middle high acidic middle middle 7 Low Low alkaline Low Low 48 middle high acidic high Low 8 Low Low alkaline middle Low 49 middle high neutral Low middle 9 Low Low alkaline high Very low 50 middle high neutral middle middle 10 Low middle acidic Low Low 51 middle high neutral high Low 11 Low middle acidic middle Low 52 middle high alkaline Low middle 12 Low middle acidic high Very low 53 middle high alkaline middle middle 13 Low middle neutral Low Low 54 middle high alkaline high Low 14 Low middle neutral middle Low 55 high Low acidic Low Medium and high 15 Low middle neutral high Very low 56 high Low acidic middle middle 16 Low middle alkaline Low Low 57 high Low acidic high middle 17 Low middle alkaline middle Low 58 high Low neutral Low high 18 Low middle alkaline high Very low 59 high Low neutral middle Medium and high 19 Low high acidic Low Very low 60 high Low neutral high middle 20 Low high acidic middle Very low 61 high Low alkaline Low Medium and high 21 Low high acidic high Very low 62 high Low alkaline middle middle 22 Low high neutral Low Very low 63 high Low alkaline high middle 23 Low high neutral middle Very low 64 high middle acidic Low Medium and high 24 Low high neutral high Very low 65 high middle acidic middle middle 25 Low high alkaline Low Very low 66 high middle acidic high middle 26 Low high alkaline middle Very low 67 high middle neutral Low high 27 Low high alkaline high Very low 68 high middle neutral middle Medium and high 28 middle Low acidic Low middle 69 high middle neutral high middle 29 middle Low acidic middle middle 70 high middle alkaline Low Medium and high 30 middle Low acidic high Low 71 high middle alkaline middle middle 31 middle Low neutral Low Medium and high 72 high middle alkaline high middle 32 middle Low neutral middle middle 73 high high acidic Low middle 33 middle Low neutral high Low 74 high high acidic middle middle 34 middle Low alkaline Low middle 75 high high acidic high middle 35 middle Low alkaline middle middle 76 high high neutral Low Medium and high 36 middle Low alkaline high Low 77 high high neutral middle middle 37 middle middle acidic Low middle 78 high high neutral high middle 38 middle middle acidic middle middle 79 high high alkaline Low middle 39 middle middle acidic high Low 80 high high alkaline middle middle 40 middle middle neutral Low Medium and high 81 high high alkaline high middle 41 middle middle neutral middle middle Normalization layer: Normalizes the activation degree of the rules.

[0051] Defuzzing layer of the wastewater parameter treatment subnetwork: The centroid method is used to convert fuzzy outputs into precise values.

[0052] The second layer of the network is a five-layer fuzzy neural network, such as... Figure 6 As shown, the structure is as follows: Input layer: 2 nodes (air treatment factor and wastewater characteristic factor); fuzzing layer: 6 nodes (3 Gaussian membership functions for each input variable); Rule layer: 9 nodes (3² = 9 fuzzy rules); Normalization layer: 9 nodes (normalizes the activation of rules); Defuzzification layer: 3 nodes (air fan speed, circulating pump speed, and concentration ratio, where the concentration ratio is the ratio of the conductivity of the concentrate to the conductivity of the desulfurization wastewater).

[0053] Specifically, the input layer and the fuzzing layer of the second-layer network: Input layer: Receives two normalized input variables: air treatment factor (range: 0-1) and wastewater characteristic factor (range: 0-1).

[0054] The fuzzification layer uses Gaussian membership functions to fuzzify the input. The rule layer defines "IF-THEN" logic rules, and each input variable has three Gaussian membership functions (low, medium, and high). The parameters of the Gaussian membership functions (center and width) can be optimized through training.

[0055] The second layer of the network consists of the rule layer and the normalization layer. Rule layer: Implements 9 fuzzy rules (3×3), as shown in Table 3 below: Table 3: Fuzzy Rules for Second-Layer Networks Serial Number Air processing factor Wastewater characteristic factors Air fan speed Circulating pump speed Concentration ratio Serial Number Air processing factor Wastewater characteristic factors Air fan speed Circulating pump speed Concentration ratio 1 Low Low high speed medium speed medium magnification 6 middle high medium speed high speed High magnification 2 Low middle high speed high speed medium magnification 7 high Low low speed low speed low magnification 3 Low high high speed high speed High magnification 8 high middle low speed medium speed medium magnification 4 middle Low medium speed low speed low magnification 9 high high medium speed medium speed High magnification 5 middle middle medium speed medium speed medium magnification Normalization layer: Normalizes the activation degree of the rules.

[0056] The second layer of the network is the deblurring layer: The centroid method is used to convert fuzzy outputs into precise values, generating optimized control parameters.

[0057] Specifically, the constructed model needs to be trained before use: based on 500 sets of laboratory simulation data and 500 sets of field operation data, the backpropagation algorithm is used to optimize the network weights, and the objective function is the weighted sum of minimizing evaporation energy consumption and maximizing evaporation efficiency. After training, the model fitting error is <5%.

[0058] S4. Feedback control and real-time adjustment.

[0059] The output value of the fuzzy neural network model is used as the setpoint for the control parameters. A PID controller is used to achieve closed-loop control, controlling the frequency converters of the air fan, wastewater circulation pump, and heater hot liquid circulation pump, as well as the operating time of the desulfurization wastewater make-up pump and concentrate discharge pump. This adjusts the air volume, circulating wastewater flow rate, circulating wastewater temperature, and concentration ratio. The PID closed-loop control logic is as follows: Airflow regulation: The deviation between the target airflow and the actual airflow is calculated in real time. The PID controller maps different airflows to different frequencies based on the airflow characteristic curve of the air fan (different speeds correspond to different airflows, which are converted to different frequencies to correspond to different airflows. The corresponding frequency for the airflow is found. For example, if the airflow of the air fan is set to 1000 m³ / h, the corresponding frequency is 30 Hz). Based on the deviation, the PID algorithm calculates the frequency adjustment of the air fan inverter. The frequency of the air fan inverter is dynamically adjusted according to the PID output adjustment, so that the actual airflow gradually approaches the target value.

[0060] Specifically, the adjustment amount of the air fan frequency is calculated so that the actual air volume gradually approaches the target value. The adjustment logic can be broken down into the following steps: Deviation calculation: Deviation value e = Target flow rate - Actual flow rate; PID Calculation: Frequency Adjustment Where Kp is the proportional coefficient, used for rapid response to deviation; Ki is the integral coefficient, used to eliminate steady-state error; and Kd is the derivative coefficient, used to suppress overshoot. Frequency adjustment: The frequency of the air blower inverter is dynamically adjusted according to the frequency adjustment amount u; Positive deviation scenario (actual air volume < target air volume): Deviation characteristic: e>0 indicates that the current air volume is insufficient and the air volume needs to be increased.

[0061] Frequency adjustment direction: Increase the PID output u → increase the fan frequency → increase the air volume, until the actual air volume is close to the target value.

[0062] For example: the target flow rate is 1000 m³ / h, the actual flow rate is 800 m³ / h (e=200), the PID calculation output u=+5Hz → the fan frequency increases from 35Hz to 40Hz, and the air volume increases.

[0063] Negative deviation scenario (actual air volume > target air volume): Deviation characteristic: e<0 indicates that the current air volume is too large and needs to be reduced.

[0064] Frequency adjustment direction: Decrease PID output u → Reduce fan frequency → Reduce air volume until the actual flow rate approaches the target value.

[0065] For example: the target flow rate is 1000 m³ / h, the actual flow rate is 1200 m³ / h (e=-200), the PID calculation output u=-5Hz → the fan frequency drops from 45Hz to 40Hz, and the air volume decreases.

[0066] Wastewater Flow Regulation: The wastewater flow rate (spray density) is controlled by adjusting the frequency of the wastewater circulation pump. Control logic: Based on the flow characteristic curve provided by the wastewater circulation pump manufacturer, different speeds correspond to different flow rates, which are then converted into different frequencies corresponding to different flow rates. The target flow rate is mapped to the frequency of the variable frequency pump. The frequency corresponding to the target flow rate is found. For example, if the flow rate is set to 1000 m³ / h, the corresponding frequency is 35 Hz, so the output is 35 Hz. The target flow rate is converted into the pump frequency, and the wastewater circulation pump frequency is dynamically corrected through a PID algorithm to achieve wastewater flow regulation.

[0067] Circulating wastewater temperature regulation: The wastewater temperature is controlled by controlling the frequency of the heater hot liquid circulation pump. The control logic is as follows: When the hot liquid temperature is fixed, the higher the frequency of the heater hot liquid circulation pump, the greater the flow rate of the hot liquid through the heat exchanger, the stronger the heat exchange capacity with the wastewater, and the higher the wastewater temperature. Conversely, the lower the pump frequency, the smaller the hot liquid flow rate, and the lower the wastewater temperature. If the hydrothermal temperature is constant at 80℃, the wastewater needs to be heated from 40℃ to different target temperatures. Temperature control is achieved by adjusting the frequency of the heater's hydrothermal circulation pump (rated range 30-50Hz). The specific adjustment process is as follows: Set the pump frequency to 30Hz, and measure that the wastewater temperature is stable at 40℃. Gradually increase the frequency to 35Hz, and the wastewater temperature rises to 45℃; continue to increase it to 40Hz, and the temperature rises to 50℃; when the frequency reaches 50Hz, the temperature stabilizes at 80℃. Record the temperature values ​​at different frequencies to form a reference for dynamic adjustment, and map it to a table corresponding to the target temperature and the frequency of the heater's hydrothermal circulation pump. The actual mapping relationship needs to be determined through on-site commissioning tests based on parameters such as the heat exchanger model, hydrothermal flow characteristics, and wastewater flow characteristics.

[0068] Concentration Ratio Adjustment: The concentration ratio is controlled by adjusting the operating time of the desulfurization wastewater make-up pump and the concentrate discharge pump. The control logic is as follows: The concentration ratio is controlled by adjusting the "make-up water" and "concentrate discharge" to change the concentration ratio. Make-up pump: The longer the operating time, the more desulfurization wastewater (make-up water) is injected, and the concentration ratio decreases. Discharge pump: The longer the operating time, the more high-concentration concentrate is discharged, and the concentration ratio increases. Based on the actual data obtained during commissioning, the concentration ratio is converted into start / stop commands for the desulfurization make-up pump and concentrate discharge pump. For example, if the current concentration ratio in the desulfurization wastewater treatment system is 3 times, it needs to be controlled between 2-5 times by adjusting the pump operating time. The commissioning process is as follows: Increase the concentration ratio to 5 times: Run the make-up pump for 10 minutes (less make-up water), run the discharge pump for 50 minutes (more concentrate discharge), and the concentration ratio increases from 3 times to 5 times. The actual operating time needs to be adjusted based on system pipelines, pump flow rates, solute concentrations, and other parameters through actual commissioning test data.

Claims

1. An optimized control method for air evaporation and concentration technology of desulfurization wastewater from thermal power plants, characterized in that: Includes the following steps: S1. Real-time acquisition of multi-source data: High-precision sensor groups are deployed inside the concentration tower and at each inlet and outlet location to collect air parameters and wastewater parameters in real time; S2. Data Preprocessing: Preprocess the collected multi-source data to construct a dataset; S3. Fuzzy Neural Network Model Construction: Based on the fuzzy neural network algorithm, a two-layer fuzzy neural network model is constructed, including a first layer network and a second layer network. The first layer network includes a parallel air parameter processing sub-network and a wastewater parameter processing sub-network, which output air treatment factors and wastewater characteristic factors, respectively. The second layer network is the decision layer, which generates air fan speed, circulation pump speed and concentration ratio based on the input air treatment factors and wastewater characteristic factors. S4. Feedback Control and Real-time Adjustment: The output value of the fuzzy neural network model is used as the control parameter setpoint. Closed-loop control is achieved through the PID controller to control the frequency conversion of the air fan, wastewater circulation pump and heater hot liquid circulation pump, as well as the running time of the desulfurization wastewater make-up pump and concentrate discharge pump, so as to adjust the air volume, circulating wastewater flow rate, circulating wastewater temperature and concentration ratio.

2. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 1, characterized in that: The air parameters collected in step S1 include air temperature, air humidity, and air flow rate. The air temperature includes the temperature at the top, middle, and bottom of the concentration tower, as well as the air inlet and outlet temperatures. The air humidity includes the humidity at the upper middle, middle, and lower middle of the concentration tower, as well as the air inlet and outlet temperatures. The air flow rate includes the flow rates at the inlet and outlet of the concentration tower. The wastewater parameters include the temperature and flow rate of the circulating wastewater, the pH value of the desulfurization wastewater and its concentrate, and the conductivity of the desulfurization wastewater and its concentrate.

3. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: In step S1, the air temperature is measured using an infrared temperature sensor with a range of 0-150℃ and an accuracy of ±1℃; the air humidity is measured using a capacitive humidity sensor with a range of 10-90%RH and an accuracy of ±1.5%; the air flow rate is measured using an ultrasonic air flow meter with a range of 0.5-20m / s and an accuracy of ±0.5%; the circulating wastewater temperature is measured using a PT100 platinum resistance temperature sensor with an accuracy of ±1℃; the circulating wastewater flow rate is measured using an electromagnetic flow meter with a range of 0.1-15m / s and an accuracy of ±0.5%; and the pH value of the desulfurization wastewater and its concentrate is measured using an online pH meter with a range of 0-14 and an accuracy of ±0.1pH. The conductivity of desulfurization wastewater and its concentrate was measured using an online conductivity meter with a measurement range of 0-300 mS / cm and an accuracy of ±1%.

4. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: The specific process of step S2 is as follows: S21. Data cleaning: Identify and remove missing values ​​and duplicate records; S22. Handling outlier data based on the 3σ principle: If a parameter exceeds the mean ± 3 times the standard deviation for 3 consecutive times, it is marked as an outlier and replaced with the mean of the previous period. S23. Correlation Analysis and Principal Component Analysis: Calculate the Pearson correlation coefficient matrix to quantify the correlation between parameters; perform principal component analysis on highly collinear features, and retain more than 95% of the variance contribution rate after dimensionality reduction; S24. Data normalization processing: Z-score standardization is used to eliminate dimensional differences. The formula is: in, The output value is the standardized value; σ represents the single original value to be standardized; μ represents the mean of the parameter dataset; σ represents the standard deviation of the parameter dataset. S25. Construct the dataset.

5. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: The air parameter processing subnetwork in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 3 nodes, each node's input variables correspond one-to-one with air humidity, air temperature and air flow; The fuzzing layer has 9 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 27 nodes, implementing 3×3×3 fuzzy rules; Normalization layer: 27 nodes, used to normalize the activation degree of rules; Deblurring layer: 1 node, which is the air processing factor. It uses the centroid method to convert the fuzzy output into an accurate value.

6. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: The wastewater parameter treatment subnetwork in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 4 nodes, each node has one-to-one input variables: circulating wastewater temperature, circulating wastewater flow rate, pH value of desulfurization wastewater and its concentrate, and conductivity of desulfurization wastewater and its concentrate; The fuzzing layer has 12 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 81 nodes, implementing 3×3×3×3 fuzzy rules; Normalization layer: 81 nodes, used to normalize the activation degree of rules; Defuzzification layer: 1 node, representing wastewater characteristic factors, using the centroid method to convert fuzzy outputs into precise values.

7. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: The second layer network in step S3 is a five-layer fuzzy neural network, with the following structure: Input layer: 2 nodes, each node's input variables correspond one-to-one with air treatment factor and wastewater characteristic factor; The fuzzing layer has 6 nodes, corresponding to 3 Gaussian membership functions for each input variable, and the parameters of the Gaussian membership functions can be optimized through training. Rule layer: 9 nodes, implementing 3×3 fuzzy rules; Normalization layer: 9 nodes, which normalizes the activation degree of rules; Defuzzification layer: 3 nodes, using the centroid method to convert fuzzy outputs into precise values, with each node's output variable corresponding to the air fan speed, circulating pump speed, and concentration ratio; the concentration ratio is the ratio of the conductivity of the concentrate to the conductivity of the desulfurization wastewater.

8. The optimized control method for air evaporation and concentration technology of desulfurization wastewater in thermal power plants according to claim 2, characterized in that: The PID closed-loop control logic in step S4 includes: Airflow regulation: Real-time calculation of the deviation between the target airflow and the actual airflow; the PID controller maps different airflows to different frequencies according to the airflow characteristic curve of the air fan, and calculates the air fan frequency adjustment amount based on the deviation value using the PID algorithm; the air fan frequency is dynamically adjusted according to the PID output adjustment amount, so that the actual airflow gradually approaches the target value. Circulating wastewater flow regulation: Based on the flow characteristic curve of the wastewater circulating pump, the target flow rate is mapped to the frequency of the variable frequency, and the frequency of the wastewater circulating pump is dynamically corrected through the PID algorithm to achieve the regulation of the circulating wastewater flow rate. Circulating wastewater temperature regulation: Record the temperature values ​​of the heater hot liquid circulation pump at different frequencies to form a reference for dynamic regulation. Map this to a table showing the correspondence between the target temperature and the frequency of the heater hot liquid circulation pump. The circulating wastewater temperature is regulated by controlling the heat exchange flow rate through the frequency of the heater hot liquid circulation pump. Concentration ratio adjustment: Based on the actual data obtained during commissioning, the concentration ratio is converted into start and stop commands for the desulfurization wastewater make-up pump and the concentrate discharge pump. The concentration ratio is adjusted by controlling the running time of the desulfurization wastewater make-up pump and the concentrate discharge pump.

Images