A wet desulfurization oxidation fan intelligent control method, device, equipment and medium

Abstract

This application discloses an intelligent control method, device, equipment, and medium for a wet desulfurization oxidation blower, relating to the field of computer technology. The method includes: collecting and utilizing the characteristics of slurry images, settling performance data, and chemical state data of the wet desulfurization system to construct an oxidation deficiency index; determining the calcium sulfite load using chemical state data and slurry circulation volume, and determining the oxygen supply matching coefficient with the oxidation deficiency index and the operating parameters of the wet desulfurization oxidation blower; generating control actions based on a pre-trained oxidation blower control strategy model and a current state vector including the oxygen supply matching coefficient, and processing the data using constraints to obtain the target action; judging whether the conditions are met based on confidence level, wet desulfurization system data quality information, and the risk of the target action, and executing the target action when the conditions are met. This application improves the stability, accuracy, and adaptability of wet desulfurization oxidation blower control while ensuring sufficient oxidation of calcium sulfite and stable slurry settling performance.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to an intelligent control method, device, equipment and medium for wet desulfurization oxidation fans. Background Technology

[0002] In wet desulfurization processes, the role of the oxidation blower is to supplement oxygen into the slurry, enabling calcium sulfite to be fully oxidized to calcium sulfate. However, a higher oxidation air volume is not always better. If the air volume is too low, calcium sulfite oxidation will be insufficient; if the air volume is too high, it will increase the blower's power consumption, enhance slurry turbulence, increase foaming or liquid level fluctuations, and may affect the operational stability of the absorption tower.

[0003] In existing oxidation blower control schemes, the following control methods mainly exist: 1) Using fixed air volume, fixed frequency, or fixed start-stop combination operation, although simple to implement, has poor adaptability. 2) Adjusting the oxidation blower based on oxidation-reduction potential, dissolved oxygen, or empirical test results, although it has some automatic adjustment capability compared to the former, still suffers from biased judgment and difficulty in timely reflecting rapidly changing load conditions. 3) Manual control, when the flue gas load changes rapidly or the slurry properties deteriorate slowly, manual adjustment is prone to problems such as untimely response, excessive or insufficient oxygen supplementation, and is difficult to trace. 4) Simply introducing intelligent models or data-driven optimization methods, because the wet desulfurization oxidation process has the characteristics of hysteresis, nonlinearity, and strong safety constraints, if the model is directly explored online in the actual desulfurization system, it may output excessive blower adjustment actions, which may lead to adverse effects. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide an intelligent control method, device, equipment, and medium for wet desulfurization oxidation blowers, which can effectively solve the problems existing in existing related solutions. While ensuring sufficient oxidation of calcium sulfite and stable slurry settling performance, it improves the stability, accuracy, and adaptability of wet desulfurization oxidation blower control, thereby reducing the ineffective energy consumption of the oxidation blower and improving the operational stability of the wet desulfurization system. The specific solution is as follows: In a first aspect, this application provides an intelligent control method for a wet desulfurization oxidation fan, including: Based on a preset data acquisition cycle, images of the slurry, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system are collected; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density; Feature extraction is performed on the slurry image to obtain slurry image features, and an oxidation deficiency index is constructed based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry; The calcium sulfite load is determined using the chemical state data and slurry circulation volume, and an oxygen supply matching coefficient is determined based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation blower; the oxygen supply matching coefficient is used to determine whether there is a current need for supplemental oxygen. Candidate control actions are generated based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data. The candidate control actions are processed using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothness constraints to obtain the target control action; Based on the confidence level of the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action, it is determined whether the preset intelligent control conditions are met. If so, the oxidation fan in the wet desulfurization system is controlled by the target control action.

[0005] Optionally, the step of extracting features from the slurry image to obtain slurry image features includes: The slurry image is preprocessed to determine a standardized image; Feature extraction is performed on the standardized image to determine the slurry image features; the slurry image features include color features, grayscale features, texture features, and foam or suspended matter features.

[0006] Optionally, the step of constructing the oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features includes: The settling performance index is determined based on the turbidity of the clear liquid layer, the settling velocity, and the height of the clear liquid layer in the settling performance data. The residual calcium sulfite factor is determined based on the calcium sulfite concentration and the slurry density. Visual anomaly factors are determined based on the similarity between the slurry image features and the normal slurry image feature library; The sedimentation performance index, the calcium sulfite residue factor, and the visual abnormality factor are weighted and fused to determine the oxidation deficiency index.

[0007] Optionally, determining the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation fan includes: The theoretical oxygen supplementation requirement is determined based on the calcium sulfite load and oxygen demand conversion factor. The actual oxygen supplementation requirement is determined based on the theoretical oxygen supplementation requirement, the oxidation deficiency index, and the oxygen supply correction coefficient. The oxygen supply capacity of the blower is determined based on the gas-liquid mass transfer correction coefficient and the blower efficiency corresponding to the wet desulfurization oxidation blower. The actual oxygen supply requirement is compared with the oxygen supply capacity of the fan to obtain the oxygen supply matching coefficient.

[0008] Optionally, generating candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model includes: Obtain an oxidation fan control strategy model that has been pre-trained offline based on the action advantages of historical control actions; wherein, the action advantages are used to characterize the superiority or inferiority of the historical control actions relative to the average action or benchmark action under the same state; Obtain the quality identifiers of the measuring points corresponding to the slurry image, the sedimentation performance data, the chemical state data, and the equipment operation data; The current state vector is determined based on the oxidation deficiency index, the oxygen supply matching coefficient, the equipment operation data, and the measurement point quality identifier. The current state vector is input into the oxidation fan control strategy model to obtain several candidate control actions; the candidate control actions include one or more of the following: fan frequency increment, air volume increment, and valve opening increment.

[0009] Optionally, the process of processing the candidate control actions using current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothness constraints to obtain the target control action includes: Determine the upper and lower limits of oxidation demand using the current operating conditions' oxidation demand constraints; The equipment safety boundary is determined using the equipment safety constraints under the current operating conditions; the equipment safety constraints include the frequency limit, current limit, and start-stop protection time of the wet desulfurization oxidation fan. Determine the slurry disturbance risk threshold using the slurry disturbance risk constraints under the current working conditions; Determine the motion change rate limit using motion smoothing constraints under the current operating conditions; The set of safe actions under the current operating condition is determined based on the upper limit of oxidation demand, the lower limit of oxidation demand, the equipment safety boundary, the slurry disturbance risk threshold, and the action change rate limit. The candidate control actions are projected onto the set of safety actions to obtain the target control action.

[0010] Optionally, the step of determining whether the preset intelligent control conditions are met based on the confidence level corresponding to the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action includes: The confidence level of the oxidation fan control strategy model is determined based on the current state vector, the candidate control action, and the pre-set confidence evaluation model. A data quality score is determined based on the clarity of the slurry image, the sensor status of the sedimentation performance data, and the degree of measurement deviation of the chemical state data. The short-term effects of the target control action after its execution are predicted, and a risk assessment is performed based on the prediction results to determine the action risk assessment value. Determine whether the confidence level is greater than a first preset threshold to determine the confidence level judgment result; Determine whether the data quality score is greater than a second preset threshold to determine the quality judgment result; Determine whether the risk assessment value of the action is greater than a third preset threshold to determine the risk assessment result; Based on the confidence level judgment result, the quality judgment result, and the risk judgment result, it is determined whether the preset intelligent control conditions are met.

[0011] Secondly, this application provides an intelligent control device for a wet desulfurization oxidation fan, comprising: The system data acquisition module is used to acquire slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system based on a preset data acquisition cycle; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density; An index construction module is used to extract features from the slurry image to obtain slurry image features, and to construct an oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry; The matching coefficient determination module is used to determine the calcium sulfite load using the chemical state data and slurry circulation volume, and to determine the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation blower; the oxygen supply matching coefficient is used to determine whether there is a current oxygen supplementation requirement; The control action generation module is used to generate candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data. The action constraint module is used to process the candidate control actions using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action. The oxidation fan control module is used to determine whether the preset intelligent control conditions are met based on the confidence level corresponding to the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action. When the conditions are met, the module uses the target control action to control the oxidation fan in the wet desulfurization system.

[0012] Thirdly, this application provides an electronic device, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the aforementioned intelligent control method for wet desulfurization oxidation fans.

[0013] Fourthly, this application provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the steps of the aforementioned intelligent control method for wet desulfurization oxidation fans.

[0014] As can be seen, in this application, slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system are collected based on a preset data acquisition cycle. The settling performance data includes the turbidity of the clear liquid layer and the settling velocity. The chemical state data includes the calcium sulfite concentration and the slurry density. Feature extraction is performed on the slurry images to obtain slurry image features, and an oxidation deficiency index is constructed based on the settling performance data, the chemical state data, and the slurry image features. The oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of settling performance abnormality in the slurry. The calcium sulfite load is determined using the chemical state data and the slurry circulation volume, and the oxygen supply is determined based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation blower. The matching coefficient is used to determine whether there is a current oxygen supply requirement. Candidate control actions are generated based on the current state vector and the pre-trained oxidation fan control strategy model. The current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operating data. The candidate control actions are processed using current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action. Based on the confidence level corresponding to the oxidation fan control strategy model, the wet desulfurization system data quality information, and the action risk assessment value corresponding to the target control action, it is determined whether the current conditions meet the preset intelligent control conditions. If so, the oxidation fan in the wet desulfurization system is controlled using the target control action.

[0015] In other words, this application first collects slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system based on a preset data acquisition cycle; then, it extracts slurry image features from the slurry images and constructs an oxidation deficiency index by combining the settling performance data and chemical state data; subsequently, it determines the calcium sulfite load using the chemical state data and slurry circulation volume, and determines the oxygen supply matching coefficient by combining the oxidation deficiency index and the operating parameters of the wet desulfurization oxidation blower; then, it inputs the current state vector, including the oxidation deficiency index, oxygen supply matching coefficient, and equipment operation data, into the pre-trained oxidation blower control strategy model to obtain candidate control actions; it processes the candidate control actions using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action; based on the confidence level of the oxidation blower control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value of the target control action, it determines whether the preset intelligent control conditions are met, and if so, executes the target control action to control the wet desulfurization oxidation blower. This approach effectively addresses the problems existing in current solutions. While ensuring sufficient oxidation of calcium sulfite and stable slurry settling performance, it improves the stability, accuracy, and adaptability of wet desulfurization oxidation fan control, thereby reducing ineffective energy consumption of the oxidation fan and enhancing the operational stability of the wet desulfurization system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 A flowchart of an intelligent control method for a wet desulfurization oxidation fan provided in this application; Figure 2 A schematic diagram of the structure of an intelligent control device for a wet desulfurization oxidation fan provided in this application; Figure 3 This application provides a structural diagram of an electronic device. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In existing oxidation blower control schemes, the following control methods mainly exist: 1) Using fixed air volume, fixed frequency, or fixed start-stop combination operation, although simple to implement, has poor adaptability. 2) Adjusting the oxidation blower based on oxidation-reduction potential, dissolved oxygen, or empirical test results, although it has some automatic adjustment capability compared to the former, still suffers from biased judgment and difficulty in timely reflecting rapidly changing load conditions. 3) Manual control, when the flue gas load changes rapidly or the slurry properties deteriorate slowly, manual adjustment is prone to problems such as untimely response, excessive or insufficient oxygen supplementation, and is difficult to trace. 4) Simply introducing intelligent models or data-driven optimization methods, because the wet desulfurization oxidation process has the characteristics of hysteresis, nonlinearity, and strong safety constraints, if the model is directly explored online in the actual desulfurization system, it may output excessive blower adjustment actions, which may lead to adverse effects.

[0020] Therefore, this application provides an intelligent control scheme for wet desulfurization oxidation blowers, which can effectively solve the problems existing in the existing related schemes. Under the premise of ensuring sufficient oxidation of calcium sulfite and stable slurry settling performance, it improves the stability, accuracy and adaptability of wet desulfurization oxidation blower control, thereby reducing the ineffective energy consumption of oxidation blowers and improving the operational stability of wet desulfurization systems.

[0021] See Figure 1 As shown in the figure, an embodiment of the present invention discloses an intelligent control method for a wet desulfurization oxidation fan, comprising: Step S11: Collect slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system based on a preset data acquisition cycle; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density.

[0022] In this embodiment, during the online operation of the wet desulfurization system with the pre-trained oxidation fan control strategy model deployed, operation-related data of the wet desulfurization system are collected periodically, including slurry images, settling performance data, chemical state data, and equipment operation data.

[0023] It is understood that this embodiment uses the safety optimization control scenario of limestone-gypsum wet desulfurization slurry oxidation blower as an example. Data such as slurry color image (i.e., slurry image), image capture time, light quality, clear liquid layer height, solid-liquid separation height, settling velocity, clear liquid layer turbidity, calcium sulfite concentration, calcium sulfate concentration, slurry density, pH, redox potential, liquid level, temperature, flue gas sulfur dioxide concentration, flue gas flow rate, circulating pump status, oxidation blower frequency, blower current, air volume, damper opening, start / stop status, and alarm status can be obtained from camera devices, settling detection devices, online instruments, PLC (Programmable Logic Controller), DCS (Distributed Control System), historical database, and manual testing system. This data provides raw operating condition data, equipment data, and quality indicators for subsequent calculations. The settling performance data includes, but is not limited to, the height of the clear liquid layer, the height of the solid-liquid separation, the settling velocity, and the turbidity of the clear liquid layer; the chemical state data includes, but is not limited to, the concentration of calcium sulfite, the concentration of calcium sulfate, the slurry density, pH, the redox potential, the liquid level, the temperature, the sulfur dioxide concentration in the flue gas, and the flue gas flow rate; and the equipment operation data includes the status of the circulating pump, the frequency of the oxidation fan, the fan current, the air volume, the damper opening, the start / stop status, and the alarm status. In addition, historical data collected from the historical database includes operating reports, laboratory data, historical trends of DCS / PLC, equipment ledgers, manual setting records, dewatering effect records, and by-product quality data.

[0024] Meanwhile, this embodiment can perform some data preprocessing operations after data acquisition, such as data alignment and quality labeling. Specifically, since slurry images, sedimentation detection, calcium sulfite analysis, online instrument data, and fan data may have different sampling periods, resampling is performed according to a unified time granularity, and time alignment relationships are established based on slurry circulation lag, oxidation reaction lag, and dewatering effect lag. Abnormal data is labeled with quality tags instead of simply being deleted. For image overexposure, insufficient lighting, lens contamination, sedimentation boundary recognition failure, missing instruments, abrupt changes, over-range values, equipment maintenance values, and communication interruption values, measurement point quality tags are generated respectively. These quality tags will subsequently be incorporated into the state vector and used as the basis for model confidence judgment. In addition, historical window features can be constructed. For key variables such as slurry color, sedimentation stratification, calcium sulfite concentration, pH, redox potential, flue gas sulfur dioxide load, air volume, and fan frequency, the current value, moving average, maximum value, minimum value, rate of change, cumulative deviation, and lag value are calculated to cover the time lag of the slurry oxidation process and sedimentation response.

[0025] It is further important to understand that the direct control object of this application can be the setpoint of the oxidation fan air volume, or the fan frequency, air valve opening, fan start / stop combination, or intermittent oxygen replenishment cycle corresponding to the oxidation air volume. The following explanation uses the oxygen replenishment amount of the oxidation fan as an example, but it is not limited to this execution method.

[0026] Step S12: Extract features from the slurry image to obtain slurry image features, and construct an oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry.

[0027] In this embodiment, the visual state of the slurry is identified based on the slurry color image and image feature vector. Normal oxidized slurry, insufficiently oxidized slurry, abnormally settling slurry, abnormally foaming slurry, and unreliable image states can be distinguished. The visual state is used as an important input for the control of the oxidation blower. An insufficient oxidation index is determined by combining settling performance data and chemical state data. Specifically: the slurry image is preprocessed to determine a standardized image; features are extracted from the standardized image to determine slurry image features; these features include color features, grayscale features, texture features, and foam or suspended matter features. A settling performance index is determined based on the turbidity of the clear liquid layer, the settling velocity, and the height of the clear liquid layer in the settling performance data; a calcium sulfite residue factor is determined based on the calcium sulfite concentration and the slurry density; a visual anomaly factor is determined based on the similarity between the slurry image features and the normal slurry image feature library; and a weighted fusion of the settling performance index, the calcium sulfite residue factor, and the visual anomaly factor is performed to determine the insufficient oxidation index.

[0028] It should be noted that in this embodiment, the slurry image undergoes processing such as region cropping, illumination correction, color space conversion, noise filtering, texture extraction, and abnormal image recognition to output slurry color vectors, grayscale features, color difference features, texture features, and foam or suspended matter features, which are then used as visual inputs for the slurry oxidation state. In one specific implementation, the visual anomaly factor can be determined by the distance between the image vector and the normal oxidized slurry image template. If the current slurry image deviates significantly from the normal oxidation state, and the calcium sulfite concentration is high or the sedimentation performance is poor, the oxidation deficiency index is increased. The visual anomaly factor is: ; in, It is a visual abnormality factor; This is a template vector for a normal oxidized slurry image; The vector is a slurry image vector, obtained from slurry image features; This represents the upper limit of the normalized distance.

[0029] The slurry image vector is: ; in, The image vector is the slurry. , , The average value of the color space; , , In order, they are hue, saturation, and brightness characteristics; It represents the grayscale distribution characteristics; Texture features; It is characterized by foam or abnormal suspended matter.

[0030] It is understood that the settling performance index in this embodiment is used to determine whether the slurry has good solid-liquid separation ability and gypsum crystal state. If the settling speed slows down, the turbidity of the clear liquid layer increases, or the solid-liquid stratification is not obvious, the calcium sulfite concentration can be used to determine whether there is insufficient oxidation. The settling performance index is as follows: ; in, Settlement performance index, This refers to the height of the clear liquid layer. The total height of the settling container. The rate of change of the settling interface. The turbidity of the supernatant. This represents the upper limit of turbidity. to For weights.

[0031] In this way, this embodiment can identify the visual settling status of the slurry. Based on slurry images, clear liquid layer height, solid-liquid separation ratio, settling velocity, turbidity changes, and historical settling curves, it can determine whether the slurry has insufficient oxidation, poor crystal quality, difficulty in settling and dewatering, or abnormal slurry viscosity, and output the visual oxidation status and settling performance index. It can also output settling risk indicators for timely warning or source tracing.

[0032] Furthermore, it can be understood that this embodiment constructs an oxidation deficiency index based on calcium sulfite load, slurry image status, sedimentation performance, pH, and redox potential. A higher oxidation deficiency index indicates more pronounced residual calcium sulfite, visual abnormalities, or sedimentation abnormalities in the current slurry, necessitating consideration of increasing the amount of oxygen supplementation. The oxidation deficiency index is as follows: ; in, The oxidation deficiency index, It is a residual factor of calcium sulfite. As a visual abnormality factor, It is a redox potential deficiency factor. As a pH correction factor, to For weights.

[0033] Step S13: Determine the calcium sulfite load using the chemical state data and slurry circulation volume, and determine the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation blower; the oxygen supply matching coefficient is used to determine whether there is a current oxygen supplementation requirement.

[0034] In this embodiment, the oxygen supply demand for oxidation will be matched with the oxygen supply of the blower. The calcium sulfite load will be calculated, and the oxygen supply matching coefficient will be determined by combining the oxidation deficiency index obtained in the previous steps. That is, the theoretical oxygen supplementation demand will be determined based on the calcium sulfite load and the oxygen demand conversion factor; the actual oxygen supplementation demand will be determined based on the theoretical oxygen supplementation demand, the oxidation deficiency index, and the oxygen supply correction factor; the oxygen supply capacity of the blower will be determined based on the gas-liquid mass transfer correction factor and the blower efficiency corresponding to the wet desulfurization oxidation blower; and the actual oxygen supplementation demand will be compared with the oxygen supply capacity of the blower to obtain the oxygen supply matching coefficient.

[0035] It is understood that this embodiment estimates the current calcium sulfite load to be oxidized based on the calcium sulfite concentration, slurry circulation rate, absorber level, flue gas sulfur dioxide load, and historical reaction lag, and determines the oxidation requirement in conjunction with the calcium sulfate formation state. An example of the calcium sulfite oxidation reaction is as follows: ; This formula is used to explain the basic relationship of calcium sulfite transforming into calcium sulfate or gypsum under the action of oxygen. The specific hydration form can be modified according to the actual slurry composition and process conditions.

[0036] The expression for estimating calcium sulfite loading is shown below: ; in, for The calcium sulfite load that needs to be oxidized at all times This refers to the slurry circulation volume or equivalent slurry volume that participates in the oxidation reaction. This refers to the concentration of calcium sulfite.

[0037] The theoretical oxygen requirement, or theoretical supplemental oxygen requirement, is: ; in, This is the theoretical oxygen requirement. The oxygen demand conversion factor is determined by the calcium sulfite oxidation reaction relationship, mass transfer efficiency, and process correction factor.

[0038] Then, the actual oxygen supplementation requirement is calculated based on the theoretical oxygen demand and the oxygen deficiency index. The actual oxygen supplementation requirement depends not only on the calcium sulfite load but also on the current degree of oxygen deficiency, gas-liquid mass transfer efficiency, blower efficiency, and slurry condition. Specifically, the actual oxygen supplementation requirement is as follows: ; in, To meet actual oxygen supplementation needs, This is the oxygen supply correction factor. This represents the enhancement factor for insufficient oxidation.

[0039] Then, the effective oxygen supply capacity of the wet desulfurization oxidation blower (i.e., the blower's oxygen supply capacity) was determined based on the operating parameters of the blower. The current oxygen supply capacity represents the amount of effective oxygen the oxidation blower can provide to the slurry under current operating conditions. Specifically, the blower's oxygen supply capacity is: ; in, To ensure the effective oxygen supply capacity of the fan, This is the gas-liquid mass transfer correction coefficient. For oxidation air flow, This represents the volume fraction of oxygen in the air. This refers to the overall efficiency of the fan and mass transfer.

[0040] The actual oxygen supply demand is compared with the effective oxygen supply capacity of the blower to obtain an oxygen supply matching coefficient. This coefficient is used to determine whether the current oxygen supply from the blower is insufficient, adequate, or excessive. The oxygen supply matching coefficient is: ; in, This is the oxygen supply matching coefficient. To prevent the minimum value where the denominator is zero. When the oxygen level is below the set lower limit, it indicates insufficient oxygen supply; when it is within the reasonable range, it indicates adequate oxygen supply; and when it exceeds the set upper limit, it indicates that the oxygen supply may be excessive.

[0041] Furthermore, this embodiment can also determine the operating direction of the oxidation fan based on the oxidation deficiency index and the oxygen supply matching coefficient, serving as a basis for subsequent decisions regarding whether to increase, maintain, or decrease the oxidation air volume. When the oxidation deficiency index is high and the oxygen supply matching coefficient is below the lower limit, the oxidation fan frequency or air volume is allowed to be increased; when the oxygen supply matching coefficient is within a reasonable range, it is maintained or slightly adjusted; when the oxygen supply matching coefficient is above the upper limit, the calcium sulfite concentration is low, or the energy consumption risk is high, the oxidation air volume is restricted or reduced. The determination of the oxidation fan operating direction is as follows: ; in, This indicates the direction of the oxidation fan's operation. The high threshold for insufficient oxidation The oxidation threshold is insufficient. and These are the lower and upper limits of the oxygen supply matching coefficient, respectively.

[0042] Step S14: Generate candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data.

[0043] In this embodiment, a state vector is constructed based on the data obtained in the preceding steps, and the constructed state vector is input into the model to generate actions for controlling the oxidation fan. That is: obtaining an oxidation fan control strategy model that has been pre-trained offline based on the action advantages of historical control actions; wherein, the action advantages are used to characterize the superiority or inferiority of the historical control actions relative to the average action or benchmark action under the same state; obtaining the measurement point quality identifiers corresponding to the slurry image, the sedimentation performance data, the chemical state data, and the equipment operation data; determining the current state vector based on the oxidation deficiency index, the oxygen supply matching coefficient, the equipment operation data, and the measurement point quality identifiers; inputting the current state vector into the oxidation fan control strategy model to obtain several candidate control actions; the candidate control actions include one or more of the following: fan frequency increment, air volume increment, and valve opening increment.

[0044] It should be understood that, in this embodiment, the state variable is a state vector that characterizes the current slurry oxidation state, sedimentation and dewatering state, calcium sulfite load, oxidation blower oxygen supply capacity, and equipment executable state. The state vector is: ; in, for Time-state vector For the current operating condition variable, For historical window variables, For the rate of change variable, Settlement performance index, For quality identification of measuring points, For device state variables, Recommended actions for traditional control This is a record of historical actions or projections. The historical window variable is the historical window feature constructed in step S11, and the measurement point quality identifier is the quality identifier in step S11. Equipment status variables are used to determine the feasibility of oxygen supplementation actions. These variables may include oxidation fan frequency, fan current, airflow feedback, damper opening, fan bearing temperature, vibration, outlet pressure, start / stop status, alarm status, and communication status. When the fan malfunctions, the airflow feedback is abnormal, or the outlet pressure exceeds the limit, traditional control or manual intervention should be triggered. Current operating condition variables consist of the collected raw operating condition data of the wet desulfurization system (involving slurry images, settling performance data, and chemical state data).

[0045] Furthermore, regarding the reinforcement learning process for the oxidation fan control strategy model, the historical operating data of the wet desulfurization system is organized into a sample sequence according to time order, with each sample including at least the state of the wet desulfurization system. Historical actual oxygen supplementation actions (i.e., historical control actions) The state at the next moment or after a certain time delay. The evaluation considers factors such as changes in calcium sulfite, changes in settling performance, calcium sulfate formation effect, energy consumption effect, and safety events. The action space is preferably incremental, including reducing the oxidation fan frequency, maintaining the oxidation fan frequency, slightly increasing the oxidation fan frequency, moderately increasing the oxidation fan frequency, adjusting the damper opening, switching fan combinations, or adjusting the intermittent oxygenation cycle; continuous actions can also be used, outputting increments in fan frequency, air volume, or valve opening. The reward function simultaneously considers the decrease in residual calcium sulfite, stable calcium sulfate formation, improved settling performance, maintained desulfurization efficiency, oxidation fan energy consumption, slurry disturbance, penalties for frequent equipment actions, and safety boundary penalties.

[0046] The reward function can be defined as follows: ; in, As a penalty for calcium sulfite residue, As a penalty for poor settling performance, Penalties for desulfurization efficiency or export sulfur dioxide risk. Due to the energy consumption penalty of the oxidation fan, Penalties for violating safety regulations Punishment based on the frequency of the action. to For weights.

[0047] It should be noted that this embodiment uses action advantage for offline model training. For historical control actions... The advantage value of an action relative to the average action value under the same conditions is calculated based on a value model. A positive advantage value indicates that the action has better oxidation, sedimentation, energy consumption, or safety effects under similar slurry visual sedimentation conditions and calcium sulfite loads; a negative advantage value indicates that the action is relatively inferior. The model prioritizes learning high-advantage actions and reduces the degree of imitation of low-advantage actions. The action advantage is defined as follows: ; in, Indicates the state Execute action The value of the action, Representing state The average or benchmark value below, This is an advantage in terms of movement.

[0048] Furthermore, to prevent the model from learning dangerous actions outside of historical operating data, conservative offline training can be performed. This involves introducing behavioral policy constraints, action range constraints, wind turbine equipment constraints, slurry disturbance penalties, and uncertainty penalties during training, so that the policy output is as close as possible to the feasible control range verified in historical data.

[0049] Furthermore, while training the strategy model offline, a confidence model can also be trained simultaneously to determine the usability of reinforcement learning actions generated by the oxidation fan control strategy model online. The confidence level can be determined based on factors such as the distance between the input state and historical data distribution, image quality, settlement detection reliability, model ensemble prediction variance, measurement point quality, number of historical similar samples, and candidate control action risk scores, generating the confidence level of the strategy model. After training, this model, along with the oxidation fan control strategy model and value model, is output as the base model. Before the initial deployment, the base model is trained only on offline data, without unconstrained exploration on-site, thus avoiding deployment risks.

[0050] Furthermore, regarding model optimization after online deployment, this embodiment stores the status, slurry image vector, settling performance index, calcium sulfite load, oxidation deficiency index, oxygen supply matching coefficient, candidate control actions, safety actions, final actions, confidence level, connection reasons, slurry response, settling response, energy consumption response, and alarm information for each control cycle after deployment as online samples. For time periods experiencing abnormal image quality, instrument failure, equipment failure, forced manual intervention, abnormal emissions, or abnormal shutdowns, the system identifies these periods to prevent them from being directly used as high-quality samples for model updates. Moreover, online optimization preferably employs periodic offline retraining or small-step updates at the edge, disallowing unconstrained random exploration of the model on-site. New samples are only used as positive samples to enhance the model if they meet the conditions of improved slurry oxidation effect, no violation of safety constraints, reliable images and measuring points, and traceable action effects. In addition, the oxidation deficiency index threshold, calcium sulfite concentration threshold, upper and lower limits of the oxygen supply matching coefficient, upper limit of liquid level fluctuation, upper limit of fan action amplitude, and confidence level threshold can be calibrated based on long-term operating data, allowing the safe and feasible domain to gradually adapt to the actual operating conditions of the specific desulfurization system. It also provides daily, weekly, or monthly statistics on the percentage of intelligent optimization actions, the number of traditional takeovers, the average frequency of oxidation fans, the power consumption of oxidation fans per unit volume of slurry, changes in calcium sulfite concentration, changes in settling performance, dewatering effect, duration of liquid level fluctuations, number of risk interceptions, and number of equipment alarms, for operators to manage or trace the source of the problem.

[0051] Step S15: Process the candidate control actions using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action.

[0052] In this embodiment, before executing an action to control the oxidation blower, it is necessary to apply safety constraints and projections to the candidate actions to ensure their safety. Specifically: The upper and lower limits of oxidation demand are determined using the current operating condition's oxidation demand constraints; the equipment safety boundary is determined using the current operating condition's equipment safety constraints; these constraints include frequency limits, current limits, and start / stop protection times for the wet desulfurization oxidation blower; the slurry disturbance risk threshold is determined using the current operating condition's slurry disturbance risk constraints; the action change rate limit is determined using the current operating condition's action smoothing constraints; a set of safe actions under the current operating condition is determined based on the upper and lower limits of oxidation demand, the equipment safety boundary, the slurry disturbance risk threshold, and the action change rate limit; and the candidate control actions are projected onto the set of safe actions to obtain the target control action. It is understood that if an action exceeds the current set of safe actions, it is projected onto the nearest safe action in the set; if no action satisfies the safety constraints, a hold action or a traditional control suggestion action is output.

[0053] The expression for motion safety projection is: ; in, Candidate control actions generated for the oxidation fan control strategy model. This is a set of safe actions under the current working conditions. This refers to safety procedures after projection.

[0054] It is important to understand that the purpose of the set of safety actions is to limit the candidate control actions output by reinforcement learning models, expert rule models, or historical similar working condition models to the range allowed by the current working condition. These actions may include, but are not limited to, safety constraints such as oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints.

[0055] In one specific implementation, the various constraint information in the safety constraints may include: oxidation demand constraints, used to ensure that the oxygen supplementation is only allowed to increase when there is insufficient oxidation or insufficient oxygen supply. When the insufficient oxidation index is below the low threshold, the calcium sulfite concentration is low, or the oxygen supply matching coefficient is above the upper limit, further increases in airflow are prohibited, or only maintaining or reducing the airflow is allowed. Liquid level and pressure constraints, used to avoid excessive disturbance within the tower after executing oxidation fan control actions, these constraints may include upper limits for liquid level and pressure. Upper and lower limit constraints for fan frequency and airflow, used to prevent insufficient oxidation due to excessively low oxidation fan operation, or increased energy consumption and equipment risk due to excessively high operation, these constraints may include upper and lower limits corresponding to fan frequency and airflow, respectively. Single action amplitude constraints, used to limit the amplitude of single adjustments, to avoid sudden changes in fan frequency, airflow surges, and control oscillations, these constraints may include upper limits for the difference in fan frequency change and airflow change between adjacent single actions. Equipment safety constraints ensure that actions can be performed by field equipment; fan current, outlet pressure, vibration, and bearing temperature should be within allowable ranges. When a fan alarms, dampers jam, or airflow feedback is abnormal, intelligent optimized actions are not allowed to directly control field equipment. Action interval constraints prevent frequent adjustments of oxidation fans or dampers; these constraints may include a lower limit for action execution intervals. Measurement point quality constraints ensure the reliability of critical input data; when critical measurement points such as slurry images, sedimentation detection, calcium sulfite concentration, pH, redox potential, airflow, or fan feedback are missing, out of range, have not been updated for a long time, or have abnormal quality indicators, reinforcement learning actions are not allowed to directly control field equipment.

[0056] Understandably, the set of safety actions is not fixed but changes dynamically based on the current oxidation state of the slurry. For example, when the calcium sulfite concentration is high, the insufficiency index is high, the settling performance is poor, and the fan equipment is in normal condition, the range of actions in the direction of increasing airflow is allowed to be appropriately increased; when the calcium sulfite concentration is low, the oxygen supply matching coefficient is high, the fan energy consumption is high, or the risk of liquid level fluctuation is high, the range of actions in the direction of increasing airflow is narrowed or shut off.

[0057] Step S16: Based on the confidence level of the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action, determine whether the preset intelligent control conditions are met. If so, use the target control action to control the oxidation fan in the wet desulfurization system.

[0058] In this embodiment, even if the action is determined to be safe, it is still necessary to determine whether the current working condition falls within the model's reliable range based on model confidence, action risk, and measurement point quality. If it does, intelligent control is implemented based on the target control action. That is: the confidence level of the oxidation blower control strategy model is determined based on the current state vector, the candidate control action, and the preset confidence evaluation model; the data quality score is determined based on the clarity of the slurry image, the sensor status of the sedimentation performance data, and the measurement deviation of the chemical state data; the short-term effect prediction after the target control action is performed, and a risk assessment is conducted based on the corresponding prediction results to determine the action risk assessment value; it is determined whether the confidence level is greater than a first preset threshold to determine the confidence level judgment result; it is determined whether the data quality score is greater than a second preset threshold to determine the quality judgment result; it is determined whether the action risk assessment value is greater than a third preset threshold to determine the risk judgment result; and based on the confidence level judgment result, the quality judgment result, and the risk judgment result, it is determined whether the current condition meets the preset intelligent control conditions.

[0059] Understandably, if any of the following factors—model confidence, action risk, or measurement point quality—fail, the current situation is deemed outside the model's confidence range, and traditional control or manual intervention is initiated. Traditional control recommendations are then used to ensure the safety of the work site. When the slurry image is overexposed, there is insufficient lighting, lens contamination, failure to identify settling boundaries, abnormal calcium sulfite measurement points, abnormal airflow feedback, or communication interruption; when the current state is far removed from historical training data distribution / the model integration output has significant discrepancies; when candidate control actions lead to excessive predicted liquid level fluctuations, excessive tower pressure, excessive fan current, excessive energy consumption, or high slurry disturbance risk; when calcium sulfite concentration continuously increases, settling performance continues to deteriorate and the model actions fail to improve the trend; or when there is an oxidation fan malfunction, abnormal current, damper jamming, abnormal outlet pressure, high bearing temperature, or abnormal frequency feedback, any of the above situations will trigger exit from intelligent control based on the oxidation fan control strategy model, and traditional control will be adopted. Traditional control methods can include fixed airflow control, segmented rule control, proportional-integral-derivative control, feedback control based on calcium sulfite concentration or redox potential, manual setpoint control, or a combination of the above. Traditional control takeover does not mean the end of learning. This embodiment also records the state before and after takeover, candidate control actions, takeover reasons, and final control effects for subsequent offline review. If a certain type of operating condition triggers takeover repeatedly, this type of operating condition can be added to the retraining sample, or a separate conservative strategy boundary can be set.

[0060] Furthermore, in this embodiment, when performing the aforementioned reliability assessment, the final control action selection can be as follows: ; in, For the final control action to be executed, For control actions after safe projection, Recommended actions for traditional control For model confidence, Due to the risk of action, The confidence threshold for the model. As the action risk threshold, This refers to the quality identification of key measuring points. Key measuring points include those related to slurry images, settling performance data, chemical state data, and equipment operation data.

[0061] In summary, this embodiment provides a safety optimization control scheme for wet desulfurization oxidation fans based on matching the visual settling state of the slurry with the calcium sulfite load. This scheme can be deployed in the host computer, edge computing device, industrial server, DCS / PLC control system, or cloud-edge collaborative platform of the flue gas treatment system. The on-site image acquisition device, online instruments, and controller are responsible for acquiring slurry images and operating data and executing fan control actions. The edge computing device or industrial server is responsible for image processing, state construction, model reasoning, safety judgment, and action recording. The PLC or DCS is responsible for the final action issuance and equipment interlock protection.

[0062] Furthermore, in this scheme: 1) the oxidation fan control based on the visual settling state of the slurry and the calcium sulfite load is not simply based on a fixed air volume or oxidation-reduction potential to control the fan, but rather on a comprehensive assessment of the slurry color image, settling performance, solid-liquid stratification state and calcium sulfite concentration to determine whether there is insufficient oxidation in the current slurry.

[0063] 2) By using slurry image vectors as input for oxidation state recognition, changes in slurry appearance can be identified through color, grayscale, texture, foam, suspended matter, and image quality indicators, so that oxidation blower control no longer relies solely on lagging test data or a single instrument.

[0064] 3) A linkage judgment mechanism between sedimentation performance and insufficient oxidation was proposed, which incorporates the height of the clear liquid layer, the solid-liquid separation ratio, the sedimentation rate and the turbidity of the supernatant into the state judgment, so that the oxygen supplementation control can take into account both the degree of calcium sulfite oxidation and the sedimentation and dehydration performance of gypsum.

[0065] 4) A matching mechanism between calcium sulfite load and oxygen supply capacity of oxidation fan was established. The actual oxygen supplementation demand was calculated based on the calcium sulfite load, and then compared with the effective oxygen supply capacity of the fan to determine whether the oxidation air volume can be increased, maintained, or decreased.

[0066] 5) Offline reinforcement learning training based on action advantages: The model does not explore without constraints on site, but learns the optimal oxygenation action under different visual sedimentation states of slurry in historical industrial data, which improves the safety and usability of the model before it goes online.

[0067] 6) Combine candidate oxygen replenishment actions with safety projection. The oxidation blower control strategy model only outputs candidate control actions. These candidate control actions must pass the constraints of oxidation demand, blower equipment, liquid level and pressure, slurry disturbance risk, and action smoothing before they can be issued.

[0068] 7) A low-confidence traditional takeover mechanism is set up. When the image quality is abnormal, the sedimentation detection is unreliable, the calcium sulfite measuring point is abnormal, the equipment is abnormal, or the model confidence is insufficient, it will automatically switch to the traditional control mode.

[0069] In this way, while ensuring that the control actions are available on-site, the risk of one-sided judgment due to single-variable feedback is avoided, and the risk of the model going online is reduced.

[0070] Therefore, in this embodiment, the following steps are taken: First, slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system are collected based on a preset data acquisition cycle. Then, slurry image features are extracted from the slurry images, and an oxidation deficiency index is constructed by combining the settling performance data and chemical state data. Next, the calcium sulfite load is determined using the chemical state data and slurry circulation volume, and the oxygen supply matching coefficient is determined by combining the oxidation deficiency index and the operating parameters of the wet desulfurization oxidation blower. Then, the current state vector, including the oxidation deficiency index, oxygen supply matching coefficient, and equipment operation data, is input into the pre-trained oxidation blower control strategy model to obtain candidate control actions. The candidate control actions are processed using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action. Based on the confidence level of the oxidation blower control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value of the target control action, it is determined whether the preset intelligent control conditions are met, and if so, the target control action is executed to control the wet desulfurization oxidation blower. This approach effectively addresses the problems existing in current solutions. While ensuring sufficient oxidation of calcium sulfite and stable slurry settling performance, it improves the stability, accuracy, and adaptability of wet desulfurization oxidation fan control, thereby reducing ineffective energy consumption of the oxidation fan and enhancing the operational stability of the wet desulfurization system.

[0071] See Figure 2 As shown in the embodiments, this application also discloses an intelligent control device for a wet desulfurization oxidation fan, comprising: The system data acquisition module 11 is used to acquire slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system based on a preset data acquisition cycle; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density; The index construction module 12 is used to extract features from the slurry image to obtain slurry image features, and to construct an oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry. The matching coefficient determination module 13 is used to determine the calcium sulfite load using the chemical state data and slurry circulation volume, and to determine the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation blower; the oxygen supply matching coefficient is used to determine whether there is a current oxygen supplementation requirement; The control action generation module 14 is used to generate candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data. The action constraint module 15 is used to process the candidate control actions using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints and action smoothing constraints to obtain the target control action. The oxidation fan control module 16 is used to determine whether the preset intelligent control conditions are met based on the confidence level corresponding to the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action. When the conditions are met, the oxidation fan in the wet desulfurization system is controlled by the target control action.

[0072] For more detailed information on the working process of each of the above modules, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0073] Furthermore, embodiments of this application also disclose an electronic device, Figure 3 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0074] Figure 3 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the intelligent control method for wet desulfurization oxidation fans disclosed in any of the foregoing embodiments. Alternatively, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0075] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0076] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0077] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the intelligent control method for the wet desulfurization oxidation fan executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0078] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned intelligent control method for wet desulfurization oxidation fans. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0079] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0080] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0081] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0082] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0083] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A wet desulphurization oxidation fan intelligent control method, characterized in that, include: Based on a preset data acquisition cycle, images of the slurry, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system are collected; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density; Feature extraction is performed on the slurry image to obtain slurry image features, and an oxidation deficiency index is constructed based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry; The calcium sulfite load is determined using the chemical state data and slurry circulation volume, and the oxygen supply matching coefficient is determined based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation fan. The oxygen supply matching coefficient is used to determine whether there is a current need for supplemental oxygen. Candidate control actions are generated based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data. The candidate control actions are processed using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothness constraints to obtain the target control action; Based on the confidence level of the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action, it is determined whether the preset intelligent control conditions are met. If so, the oxidation fan in the wet desulfurization system is controlled by the target control action.

2. The method of claim 1, wherein the method further comprises: The step of extracting features from the slurry image to obtain slurry image features includes: The slurry image is preprocessed to determine a standardized image; Feature extraction is performed on the standardized image to determine the slurry image features; the slurry image features include color features, grayscale features, texture features, and foam or suspended matter features.

3. The method of claim 1, wherein the method further comprises: The construction of the oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features includes: The sedimentation performance index is determined based on the turbidity of the clear liquid layer, the sedimentation velocity, and the height of the clear liquid layer in the sedimentation performance data. The residual calcium sulfite factor is determined based on the calcium sulfite concentration and the slurry density. Visual anomaly factors are determined based on the similarity between the slurry image features and the normal slurry image feature library; The sedimentation performance index, the calcium sulfite residue factor, and the visual abnormality factor are weighted and fused to determine the oxidation deficiency index.

4. The method of claim 1, wherein the method further comprises: The determination of the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation fan includes: The theoretical oxygen supplementation requirement is determined based on the calcium sulfite load and oxygen demand conversion factor. The actual oxygen supplementation requirement is determined based on the theoretical oxygen supplementation requirement, the oxidation deficiency index, and the oxygen supply correction coefficient. The oxygen supply capacity of the blower is determined based on the gas-liquid mass transfer correction coefficient and the blower efficiency corresponding to the wet desulfurization oxidation blower. The actual oxygen supply requirement is compared with the oxygen supply capacity of the fan to obtain the oxygen supply matching coefficient.

5. The method of claim 1, wherein the method further comprises: The generation of candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model includes: Obtain an oxidation fan control strategy model that has been pre-trained offline based on the action advantages of historical control actions; wherein, the action advantages are used to characterize the superiority or inferiority of the historical control actions relative to the average action or benchmark action under the same state; Obtain the quality identifiers of the measuring points corresponding to the slurry image, the sedimentation performance data, the chemical state data, and the equipment operation data; The current state vector is determined based on the oxidation deficiency index, the oxygen supply matching coefficient, the equipment operation data, and the measurement point quality identifier. The current state vector is input into the oxidation fan control strategy model to obtain several candidate control actions; the candidate control actions include one or more of the following: fan frequency increment, air volume increment, and valve opening increment.

6. The method of claim 1 to 5, wherein, The process of processing the candidate control actions using current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothness constraints to obtain the target control action includes: Determine the upper and lower limits of oxidation demand using the current operating conditions' oxidation demand constraints; The equipment safety boundary is determined using the equipment safety constraints under the current operating conditions; the equipment safety constraints include the frequency limit, current limit, and start-stop protection time of the wet desulfurization oxidation fan. Determine the slurry disturbance risk threshold using the slurry disturbance risk constraints under the current working conditions; Determine the motion change rate limit using motion smoothing constraints under the current operating conditions; The set of safe actions under the current operating condition is determined based on the upper limit of oxidation demand, the lower limit of oxidation demand, the equipment safety boundary, the slurry disturbance risk threshold, and the action change rate limit. The candidate control actions are projected onto the set of safety actions to obtain the target control action.

7. The method of claim 1, wherein the method further comprises: The step of determining whether the preset intelligent control conditions are met based on the confidence level of the oxidation blower control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action includes: The confidence level of the oxidation fan control strategy model is determined based on the current state vector, the candidate control action, and the pre-set confidence evaluation model. A data quality score is determined based on the clarity of the slurry image, the sensor status of the sedimentation performance data, and the degree of measurement deviation of the chemical state data. The short-term effects of the target control action after its execution are predicted, and a risk assessment is performed based on the prediction results to determine the action risk assessment value. Determine whether the confidence level is greater than a first preset threshold to determine the confidence level judgment result; Determine whether the data quality score is greater than a second preset threshold to determine the quality judgment result; Determine whether the risk assessment value of the action is greater than a third preset threshold to determine the risk assessment result; Based on the confidence level judgment result, the quality judgment result, and the risk judgment result, it is determined whether the preset intelligent control conditions are met.

8. A wet desulphurization oxidation fan intelligent control device, characterized in that, include: The system data acquisition module is used to acquire slurry images, settling performance data, chemical state data, and equipment operation data of the wet desulfurization system based on a preset data acquisition cycle; the settling performance data includes the turbidity of the clear liquid layer and the settling velocity; the chemical state data includes the calcium sulfite concentration and the slurry density; An index construction module is used to extract features from the slurry image to obtain slurry image features, and to construct an oxidation deficiency index based on the sedimentation performance data, the chemical state data, and the slurry image features; the oxidation deficiency index is used to characterize the degree of calcium sulfite residue, the degree of visual abnormality, and the degree of sedimentation performance abnormality in the slurry; The matching coefficient determination module is used to determine the calcium sulfite load using the chemical state data and slurry circulation volume, and to determine the oxygen supply matching coefficient based on the calcium sulfite load, the oxidation deficiency index, and the operating parameters of the wet desulfurization oxidation fan. The oxygen supply matching coefficient is used to determine whether there is a current need for supplemental oxygen. The control action generation module is used to generate candidate control actions based on the current state vector and the pre-trained oxidation fan control strategy model; the current state vector includes the oxidation deficiency index, the oxygen supply matching coefficient, and the equipment operation data. The action constraint module is used to process the candidate control actions using the current oxidation demand constraints, equipment safety constraints, slurry disturbance risk constraints, and action smoothing constraints to obtain the target control action. The oxidation fan control module is used to determine whether the preset intelligent control conditions are met based on the confidence level corresponding to the oxidation fan control strategy model, the data quality information of the wet desulfurization system, and the action risk assessment value corresponding to the target control action. When the conditions are met, the module uses the target control action to control the oxidation fan in the wet desulfurization system.

9. An electronic device, comprising: include: Memory, used to store computer programs; A processor for executing the computer program to implement the intelligent control method for wet desulfurization oxidation blowers as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer programs, which, when executed by a processor, implement the intelligent control method for wet desulfurization oxidation fans as described in any one of claims 1 to 7.

Images