Optimization method of coke oven gas desulfurization system based on K + / Na + ratio monitoring and self-adaptive optimization system thereof

Abstract

The invention provides an optimization method of a coke oven gas desulfurization system based on K + / Na + ratio monitoring and a self-adaptive optimization system thereof. The method comprises the following steps: acquiring ion concentration data reflecting the chemical state of a solution and data of key process parameters in the desulfurization system in real time; inputting the data collected in real time into a prediction model, wherein the prediction model outputs a predicted value of the future barren liquor H2S concentration and feedforward control parameters used for counteracting efficiency reduction; on the basis of the feedforward control parameters, a control instruction is generated and acts on an execution mechanism, and meanwhile the actual measurement value of the H2S concentration of the barren liquor is monitored; and taking a deviation value between an actual measurement value and a predicted value of the H2S concentration of the barren liquor as a feedback signal, correcting a feedforward control parameter, and using the deviation value for parameter updating of a prediction model. According to the method, the K + / Na + ratio is introduced as a monitoring index, and the prediction model and feedforward-feedback compound control are combined, so that early diagnosis and trend prediction of desulfurization efficiency reduction are realized, and accurate and adaptive optimization adjustment is performed.

Description

Technical Field

[0001] This invention relates to the technical field of coke oven gas purification, and particularly to a method based on K... + / Na + An optimization method for a coke oven gas desulfurization system based on ratio monitoring and its adaptive optimization system. Background Technology

[0002] Vacuum potassium carbonate process, as one of the mainstream technologies in the field of coke oven gas desulfurization and purification, relies on the efficient absorption of acidic gas components such as hydrogen sulfide (H2S) from the gas by potassium alkali solution. The operational stability of this process directly affects the production stability of subsequent stages and the environmental compliance of the final flue gas emissions, thus holding a pivotal position in industries such as coking and coal chemical engineering.

[0003] However, the traditional methods used in process control for this technology have long had significant shortcomings. Currently, mainstream control strategies generally rely on delayed feedback from online monitoring data of H2S concentration in the outlet gas pipeline, or on empirical adjustments to macroscopic operating parameters such as regeneration tower temperature and total system pressure based on operator experience. This "remedial" control logic is essentially a passive response and is ill-suited to the complex effects of the dynamic evolution of components in the desulfurization solution (mainly KOH solution) during long-term circulation. With prolonged operation, impurity ions such as sodium ions and thiocyanate ions accumulate in the solution, gradually deteriorating the chemical environment and relatively reducing the effective concentration of active components, thus causing a slow and subtle decline in desulfurization efficiency. Because there is a significant transmission delay and time lag between a substantial decrease in solution activity and the detection of excessive H2S concentration at the outlet, control actions based on end-concentration feedback are always "half a beat too late." This control delay not only easily causes fluctuations in production conditions, affecting the stable operation of downstream users, but also harbors a significant risk of instantaneous exceedances of environmental emission standards due to untimely response.

[0004] On the other hand, existing control methods are severely inadequate in predicting the performance degradation of desulfurization systems. Due to the lack of real-time and effective online monitoring and in-depth analysis capabilities for key indicators reflecting the intrinsic chemical state of the desulfurization liquid (such as the potassium-sodium ion concentration ratio, active potassium carbonate concentration, and by-product salt content), traditional control models can only initiate intervention measures after a significant decline in desulfurization efficiency and the full exposure of problems. This undoubtedly misses the optimal window for early fine-tuning and prevention. Furthermore, parameter adjustments largely rely on the operator's personal experience and judgment, lacking precise and quantitative decision support based on process mechanism models and real-time operating data. This relatively crude control approach often leads to excessive replenishment of potassium alkali (KOH solution), increasing chemical consumption, or blindly increasing regeneration temperature and steam consumption to maintain desulfurization efficiency, resulting in energy waste and making it difficult for the system to operate within the optimal operating range that balances economy, environmental protection, and stability.

[0005] In summary, breaking through the limitations of traditional lag control and developing an adaptive control system that deeply integrates key ion monitoring, trend prediction, and intelligent optimization algorithms based on the chemical nature of solutions, to achieve early warning of desulfurization efficiency decline and precise closed-loop optimization of operating parameters, has become an urgent industry need to improve the operation level of vacuum potassium carbonate desulfurization process and achieve green, low-carbon, cost-reduction, and efficiency improvement. Summary of the Invention

[0006] This invention is proposed to address the above-mentioned shortcomings, and aims to provide a K-based... + / Na + An optimization method for coke oven gas desulfurization systems based on ratio monitoring and its adaptive optimization system. This method introduces K... + / Na + The ratio, as a key monitoring indicator, combined with a predictive model and feedforward-feedback composite control, enables early diagnosis and trend prediction of desulfurization efficiency decline, transforming passive response into proactive intervention. It also allows for precise and adaptive optimization and adjustment of key parameters, while scientific replacement decisions are made based on real-time cost analysis, ultimately achieving the goal of efficient, stable, economical, and intelligent operation of the desulfurization system.

[0007] To achieve the above objectives, the first aspect of the present invention provides a K-based... + / Na + The optimization method for coke oven gas desulfurization system based on ratio monitoring includes the following steps: Step S1: Real-time acquisition of ion concentration data and key process parameters reflecting the chemical state of the solution in the desulfurization system; Step S2: Input the real-time collected data into the prediction model, which outputs the predicted value of the future lean liquid H2S concentration and the feedforward control parameters used to offset the decrease in efficiency. Step S3: Based on the feedforward control parameters, generate control commands and apply them to the actuator, while monitoring the actual measured value of H2S concentration in the lean solution; Step S4: Use the deviation between the actual measured value and the predicted value of H2S concentration in the lean solution as a feedback signal to correct the feedforward control parameters, and use the deviation value to update the parameters of the prediction model to form a closed-loop adaptive optimization.

[0008] Further, in step S1, the ion concentration data reflecting the chemical state of the solution are the K values ​​of potassium ion concentration and sodium ion concentration. + / Na + The ratio; the key process parameters include one or more of the following: lean liquor temperature, lean liquor circulation flow rate, regeneration tower temperature, and total system pressure.

[0009] Further, in step S2, the prediction model is a regression model trained based on the gradient boosting decision tree algorithm; the gradient boosting decision tree algorithm adopts the LightGBM framework, and its objective function is: Where L is the loss function, Ω is the regularization term, and f t The output of the model for the t-th tree is... x i For the first i The feature vector of each sample includes the core ion ratio parameter, i.e., K. + / Na + The ratio; and key process parameters, such as lean liquor temperature, lean liquor circulation flow rate, regeneration tower temperature, and total system pressure; y i For the first i The true value of each sample is the actual measured value of H2S concentration in the depleted solution; t is the iteration round number (tree index), which is the t-th tree currently being trained; For the first t−1 trees, pair the th i The cumulative predicted value of each sample is the H2S concentration predicted by the model after the first t-1 iterations.

[0010] Furthermore, in step S4, the PID correction formula for the feedforward control parameters is: Where u(t) is the correction amount for the feedforward control parameters, e(t) is the deviation between the actual measured value and the predicted value, and K p K is the proportionality coefficient. i K is the integral coefficient. d These are the differential coefficients. t This refers to the real-time duration of continuous system operation.

[0011] Furthermore, in step S4, the feedforward control parameters include the adjustment amount of the lean solution temperature setpoint and the adjustment amount of the lean solution circulation flow rate.

[0012] Furthermore, the method also includes: economic replacement decision-making, and real-time calculation of the comprehensive cost C for the continued operation of the desulfurization system. run The overall cost of solution replacement with the desulfurization system C replace When C is satisfied run >C replace This automatically triggers the replacement procedure. Furthermore, the overall cost C of continuing to operate the desulfurization system run The overall cost of solution replacement with the desulfurization system C replace They are calculated using the following formulas respectively: C run = k 1 ·E 能耗 + k 2· M KOH溶液 + k 3 ·L 效率损失 C replace =C 新液 +C 停机 in, k 1 Energy cost coefficient k 2 The cost coefficient for KOH solution. k 3 E is the efficiency loss cost coefficient. 能耗 For energy consumption costs, M KOH溶液 For the cost of KOH solution, L 效率损失 To reduce costs due to efficiency, C 新液 For the cost of new KOH solution, C 停机 Downtime costs.

[0013] A second aspect of the present invention provides an adaptive optimization system for implementing the above-described method, comprising: The data sensing unit, including an online ion concentration analyzer and process parameter sensors, is used to collect real-time operating data of the desulfurization system; The intelligent decision-making unit includes a server storing predictive models, which processes the sensed data and generates control commands. The execution control unit includes a temperature controller, a flow controller, and automatic valves that can receive commands, for performing process adjustments; The closed-loop feedback unit is used to compare the actual measured values ​​and predicted values ​​of the desulfurization system in real time, and drive the prediction model to achieve parameter self-update, thereby forming a complete intelligent optimization closed loop.

[0014] Furthermore, the intelligent decision-making unit also includes an economic optimization algorithm module. This module takes the ratio of potassium ion concentration to sodium ion concentration in the desulfurization system as input and automatically calculates the optimal solution replacement timing by comparing the long-term expected total cost under different decisions.

[0015] Furthermore, the economic optimization algorithm module calculates the optimal replacement timing based on the dynamic programming principle, and its state transition equation is: in, V(t) This represents the minimum expected total cost starting from time t. The immediate cost of choosing to continue running at time t. The immediate cost of choosing to perform solution displacement at time t. Let represent the minimum expected total cost starting from the next decision time t+1. This represents the minimum expected total cost when the system returns to its initial state after solution displacement.

[0016] Compared with the prior art, the present invention has the following beneficial effects: Firstly, this invention proposes using K + / Na + The novel concept of using the ratio as a key monitoring indicator reveals for the first time the intrinsic relationship between the potassium-sodium ion imbalance caused by sodium ion accumulation in potassium alkali solution and the decline in desulfurization efficiency. + / Na + The ratio was established as an early and sensitive key parameter for characterizing solution activity and predicting desulfurization performance.

[0017] Secondly, this invention constructs an intelligent prediction model based on ion ratio and process parameters, and establishes a system based on K... + / Na + Predictive models (such as LightGBM) that take real-time data such as ratios, temperatures, and flow rates as inputs and output future lean liquor H2S concentrations can achieve early prediction and trend warning of desulfurization efficiency.

[0018] Third, this invention designs a feedforward-feedback composite adaptive control strategy, which uses the feedforward control signal output by the predictive model to adjust operating parameters such as temperature and flow rate in advance to offset efficiency decay; at the same time, the feedback signal of the actual concentration of lean H2S is introduced to perform PID correction on the feedforward control quantity and update the model parameters online to form a closed-loop optimization, effectively overcoming the lag of traditional control.

[0019] Fourth, this invention develops an economical replacement decision mechanism based on dynamic cost analysis. By calculating the cost of continued operation and the cost of solution replacement in real time, it automatically judges and triggers the optimal replacement time based on a dynamic programming model, thereby minimizing the total life cycle operating cost while ensuring the desulfurization effect.

[0020] In summary, this invention overcomes the shortcomings of existing vacuum potassium carbonate desulfurization processes, such as response lag, extensive control, and unpredictable efficiency decline, which rely on end-of-pipe H2S concentration detection and macroscopic empirical control. This is achieved by introducing K... + / Na + The ratio, as a key monitoring indicator, combined with predictive models and feedforward-feedback composite control, enables early diagnosis and trend prediction of desulfurization efficiency decline, transforming passive response into proactive intervention. It also allows for precise and adaptive optimization of key parameters such as regeneration temperature and lean liquor circulation volume, achieving early warning and precise intervention for desulfurization efficiency decline, significantly improving the control stability of H2S concentration in the outlet gas. Intelligent prediction and feedforward-feedback control enable fine adjustment of operating parameters, effectively reducing KOH solution consumption and energy consumption. Combined with an economical replacement decision-making mechanism, scientific replacement decisions are made based on real-time cost analysis, ensuring desulfurization effectiveness while optimizing the entire lifecycle cost. Ultimately, this comprehensively improves the reliability, economy, and intelligence of the system, achieving the ultimate goal of efficient, stable, economical, and intelligent operation of the desulfurization system. Attached Figure Description

[0021] Figure 1 This invention is based on K + / Na + Control logic diagram of the optimization method for coke oven gas desulfurization system based on ratio monitoring. Detailed Implementation

[0022] The following examples illustrate the implementation of the present invention in detail, but they do not constitute a limitation on the invention and are merely illustrative. Furthermore, the advantages of the present invention will become clearer and easier to understand by explaining them.

[0023] A K-based invention + / Na + Optimization methods for coke oven gas desulfurization systems based on ratio monitoring include: Step S1: Real-time acquisition of ion concentration data and key process parameters reflecting the chemical state of the solution in the desulfurization system; the ion concentration data reflecting the chemical state of the solution is the K0 of potassium ion concentration and sodium ion concentration. + / Na + Ratio, the ratio of potassium ion concentration to sodium ion concentration K + / Na +The range is 45 to 60; key process parameters include one or more of the following: lean liquor temperature, lean liquor circulation flow rate, regeneration tower temperature, and total system pressure.

[0024] Step S2: Input the real-time collected data into the prediction model, which outputs a predicted value for the future H2S concentration in the low-sodium solution (0.6~0.8 g / m³). 3 ) and feedforward control parameters used to offset efficiency degradation; The prediction model is a regression model trained using the gradient boosting decision tree algorithm. The gradient boosting decision tree algorithm employs the LightGBM framework, with the following model parameters: learning rate of 0.1–0.2, maximum tree depth of 5–7, minimum leaf node data size of 18–22, and objective function: Where L is the loss function, Ω is the regularization term, and f t The output of the model for the t-th tree is... x i For the first i The feature vector of each sample includes: the core ion ratio parameter, i.e., K. + / Na + The ratio; and key process parameters, such as lean liquor temperature, lean liquor circulation flow rate, regeneration tower temperature, and total system pressure; y i For the first i The true value of each sample is the actual measured value of H2S concentration in the depleted solution; t is the iteration round number (tree index), which is the t-th tree currently being trained; For the first t−1 trees, pair the th i The cumulative predicted value of each sample is the H2S concentration predicted by the model after the first t-1 iterations.

[0025] Step S3: Based on the feedforward control parameters, generate control commands and apply them to the actuator, while monitoring the actual measured value of H2S concentration in the lean solution; Step S4: Use the deviation between the actual measured value and the predicted value of H2S concentration in the lean solution as a feedback signal to correct the feedforward control parameters, and use the deviation value to update the parameters of the prediction model to form a closed-loop adaptive optimization.

[0026] The feedforward control parameters include adjustments to the lean solution temperature setpoint and the lean solution circulation flow rate. The lean solution temperature setpoint ranges from 30 to 35°C, and the lean solution circulation flow rate ranges from 165 to 175 m³ / h. 3 / h. The PID correction formula for the feedforward control parameters is: Where u(t) is the correction amount for the feedforward control parameters, e(t) is the deviation between the actual measured value and the predicted value, and K p K is the proportionality coefficient. i K is the integral coefficient. d These are the differential coefficients. t This represents the real-time duration of continuous system operation. Preferably, the scaling factor K... p The integral coefficient K is between 1.5 and 2.5. i The differential coefficient K is between 0 and 0.2. d It ranges from 0.1 to 1.0.

[0027] Step S5: Economic replacement decision, real-time calculation of the comprehensive cost C of continuing to operate the desulfurization system. run The overall cost of solution replacement with the desulfurization system C replace When C is satisfied run >C replace The replacement procedure is automatically triggered. The overall cost C for the continued operation of the desulfurization system is... run The overall cost of solution replacement with the desulfurization system C replace They are calculated using the following formulas respectively: C run = k 1 ·E 能耗 + k 2· M KOH溶液 + k 3 ·L 效率损失 C replace =C 新液 +C 停机 in, k 1 Energy cost coefficient k 2 The cost coefficient for KOH solution. k 3 E is the efficiency loss cost coefficient. 能耗 For energy consumption costs, M KOH溶液 For the cost of KOH solution, L 效率损失 To reduce costs due to efficiency, C 新液 For the cost of new KOH solution, C 停机 This refers to downtime costs. Preferably, the energy cost coefficient is... k 1 The cost coefficient for KOH solution is 0.7–0.9. k 2 The efficiency loss cost coefficient is between 1.1 and 1.3. k 3 The value is 148-152.

[0028] An adaptive optimization system for implementing the above method according to the present invention includes: The data sensing unit includes an online ion concentration analyzer and process parameter sensors, used to collect real-time operating data of the desulfurization system; the online ion concentration analyzer is an online ion chromatograph used to monitor potassium ion concentration and sodium ion concentration in real time. The intelligent decision-making unit includes a server storing predictive models, which processes the sensed data and generates control commands. The execution control unit includes a temperature controller, a flow controller, and automatic valves that can receive commands, for performing process adjustments; The closed-loop feedback unit is used to compare the actual measured values ​​and predicted values ​​of the desulfurization system in real time, and drive the prediction model to achieve parameter self-update, thereby forming a complete intelligent optimization closed loop.

[0029] The intelligent decision-making unit also includes an economic optimization algorithm module. This module takes the ratio of potassium ion concentration to sodium ion concentration in the desulfurization system as input and automatically calculates the optimal solution replacement timing by comparing the long-term expected total cost under different decisions. The economic optimization algorithm module calculates the optimal replacement timing based on the principle of dynamic programming, and its state transition equation is as follows: in, V(t) This represents the minimum expected total cost starting from time t. The immediate cost of choosing to continue running at time t. The immediate cost of choosing to perform solution displacement at time t. Let represent the minimum expected total cost starting from the next decision time t+1. This represents the minimum expected total cost when the system returns to its initial state after solution displacement.

[0030] Example: This embodiment is based on K + / Na + The optimization method for coke oven gas desulfurization systems based on ratio monitoring includes the following steps: I. System Configuration and Parameter Initialization (a) Data sensing unit: 1) Install an online ion chromatograph on the lean liquor circulation pipeline of the desulfurization system to monitor potassium ions (K+) in real time. + ) and sodium ions (Na + The concentration of ) and calculate K + / Na + The ratio is updated once per hour. 2) Install temperature and flow sensors to collect real-time data on the temperature (°C) and circulation flow rate (m³) of the lean solution. 3 / h); 3) Install an online H2S analyzer after the desulfurization tower to monitor the actual H2S concentration in the lean solution in real time (unit: mg / m³). 3 ).

[0031] (II) Intelligent Decision-Making Unit: 1) An industrial server is used, with the aforementioned intelligent decision-making software built-in. The prediction model uses the LightGBM regression algorithm, and its initial model parameters are set as follows: learning rate of 0.1, maximum tree depth of 6, and minimum number of data points per leaf node of 20; 2) The model training data uses the historical operating data of the desulfurization system over the past 3 months. Feature variables include: K + / Na + The target variables were: ratio, lean solution temperature, and lean solution flow rate; the average H2S concentration in the lean solution over the next 2 hours. After training, the model's mean absolute error (e) between predicted and actual values ​​was less than 5 mg / m³. 3 ; 3) The initial parameters of the PID controller are set as follows: proportional coefficient K p =2.0, integral coefficient K i =0.1, differential coefficient K d =0.5; 4) In the economic decision-making model, the initial cost coefficients are set as follows: k1 (energy consumption cost coefficient) = 0.8, k2 (KOH solution cost coefficient) = 1.2, k3 (efficiency loss cost coefficient) = 150. The cost C of the new KOH solution... 新液 and downtime costs C 停机 Set according to current market prices.

[0032] (iii) Execution control unit: This includes a temperature controller connected to the reboiler of the regeneration tower, a frequency converter (for regulating flow) connected to the lean liquor circulation pump, and an automatic control valve on the solution displacement pipeline.

[0033] II. System Operation Process The operation process of the method of the present invention is as follows, and refers to... Figure 1 The control logic shown is as follows: (a) Data Acquisition and Input: The data sensing unit continuously collects K + / Na + The ratio (the measured value is 50, i.e., K) + :Na + =50:1), lean solution temperature (32℃), lean solution flow rate (170m³ / h) 3Real-time data such as / h) is transmitted to the intelligent decision-making unit; (II) Intelligent Prediction and Feedforward Control: The intelligent decision-making unit inputs the received real-time data into the pre-trained LightGBM prediction model. Based on the current operating conditions, the model predicts that the H2S concentration in the lean solution will reach 580 mg / m³ in the next 2 hours. 3 (Approaching the control target of 600 mg / m³) 3 (The critical value). Simultaneously, based on the prediction results, the feedforward controller calculates that to offset this efficiency decline, the preset value of the lean solution temperature needs to be increased by 1.5℃, and the lean solution circulation flow rate needs to be increased by 20m³. 3 / h. This feedforward control command is sent to the execution control unit; (III) Command Execution and Process Monitoring: After receiving the command, the execution control unit adjusts the regeneration tower temperature setpoint from 60℃ to 61.5℃, and the frequency converter adjusts the lean liquor circulation flow rate from 170m³ / h to 61.5℃. 3 / h increased to 190m 3 / h. The system operates under these new parameters. Simultaneously, the online H2S analyzer continuously monitors the actual concentration of H2S in the lean solution; (iv) Feedback correction and model update: 1) After 1 hour, the actual measured concentration of H2S in the lean solution was 573 mg / m³. 3 , compared with the model prediction (580 mg / m 3 -7mg / m 3 The deviation value (e); 2) This deviation value is fed into the PID feedback corrector. The PID controller uses the formula: Perform calculations and output a small correction value, such as fine-tuning the temperature adjustment from +1.5℃ to +1.3℃; 3) At the same time, the actual operating data containing the bias is stored in the database and used to perform incremental learning on the LightGBM prediction model every 24 hours, so as to realize online self-updating of model parameters and make the prediction more and more accurate.

[0034] (v) Economic replacement decision: 1) The economic optimization algorithm module in the intelligent decision-making unit calculates the comprehensive cost C of the desulfurization system continuing to operate in real time every day. run The overall cost of solution replacement with the desulfurization system C replace ; 2) On a certain day, calculations showed that due to the continuous decline in efficiency, the consumption of KOH solution increased significantly, leading to C... run >C replace ; 3) The economic decision-making module immediately sends an instruction to the execution control unit to automatically open the replacement valve, discharge some of the failed KOH solution and replenish it with new KOH solution, thereby minimizing the total life cycle operating cost while ensuring the desulfurization effect.

[0035] (vi) Implementation Results Compared to traditional control methods, the fluctuation range of H2S concentration in the outlet gas is significantly reduced, and the stability is improved by more than 30%. Through precise adjustment, the consumption of regenerated steam is reduced by about 8%, and the consumption of desulfurizing agent is reduced by about 12%. The system can automatically determine the optimal replacement time, avoiding waste or risks caused by human experience-based misjudgments, and achieving safe, stable, and economical intelligent operation.

[0036] The above are merely specific embodiments of the present invention. It should be noted that any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention. Any other aspects not described in detail are prior art.

Claims

1. A K-based + / Na + The method for optimizing a coke oven gas desulfurization system based on ratio monitoring is characterized by, Includes the following steps: Step S1: Real-time acquisition of ion concentration data and key process parameters reflecting the chemical state of the solution in the desulfurization system; Step S2: Input the real-time collected data into the prediction model, which outputs the predicted value of the future lean liquid H2S concentration and the feedforward control parameters used to offset the decrease in efficiency. Step S3: Based on the feedforward control parameters, generate control commands and apply them to the actuator, while monitoring the actual measured value of H2S concentration in the lean solution; Step S4: Use the deviation between the actual measured value and the predicted value of H2S concentration in the lean solution as a feedback signal to correct the feedforward control parameters, and use the deviation value to update the parameters of the prediction model to form a closed-loop adaptive optimization.

2. The method according to claim 1, characterized in that, In step S1, the ion concentration data reflecting the chemical state of the solution are the potassium ion concentration and the sodium ion concentration, K0. + / Na + The ratio; the key process parameters include one or more of the following: lean liquor temperature, lean liquor circulation flow rate, regeneration tower temperature, and total system pressure.

3. The method according to claim 1, characterized in that, In step S2, the prediction model is a regression model trained based on the gradient boosting decision tree algorithm; the gradient boosting decision tree algorithm uses the LightGBM framework, and its objective function is: Where L is the loss function, Ω is the regularization term, and f t The output of the model for the t-th tree is... x i For the first i The feature vector of each sample y i For the first i The true value of each sample, where t is the number of iterations; For the first t−1 trees, pair the th i The cumulative predicted value of each sample.

4. The method according to claim 1, 2, or 3, characterized in that, In step S4, the PID correction formula for the feedforward control parameters is: Where u(t) is the correction amount for the feedforward control parameters, e(t) is the deviation between the actual measured value and the predicted value, and K p K is the proportionality coefficient. i K is the integral coefficient. d These are the differential coefficients. t This refers to the real-time duration of continuous system operation.

5. The method according to claim 4, characterized in that, In step S4, the feedforward control parameters include the adjustment amount of the lean solution temperature setpoint and the adjustment amount of the lean solution circulation flow rate.

6. The method according to claim 1, 2, or 3, characterized in that, Also includes: Economic replacement decision-making, real-time calculation of the comprehensive cost C of continuing to operate the desulfurization system. run The overall cost of solution replacement with the desulfurization system C replace When C is satisfied run >C replace The replacement procedure is automatically triggered.

7. The method according to claim 6, characterized in that, The overall cost C for the continued operation of the desulfurization system run The overall cost of solution replacement with the desulfurization system C replace They are calculated using the following formulas respectively: C run = k 1 ·E 能耗 + k 2· M KOH溶液 + k 3 ·L 效率损失 C replace =C 新液 +C 停机 in, k 1 Energy cost coefficient k 2 The cost coefficient for KOH solution. k 3 E is the efficiency loss cost coefficient. 能耗 For energy consumption costs, M KOH溶液 For the cost of KOH solution, L 效率损失 To reduce costs due to efficiency, C 新液 For the cost of new KOH solution, C 停机 Downtime costs.

8. An adaptive optimization system for implementing the method according to any one of claims 1 to 7, characterized in that, include: The data sensing unit, including an online ion concentration analyzer and process parameter sensors, is used to collect real-time operating data of the desulfurization system; The intelligent decision-making unit includes a server storing predictive models, which processes the sensed data and generates control commands. The execution control unit includes a temperature controller, a flow controller, and automatic valves that can receive commands, for performing process adjustments; The closed-loop feedback unit is used to compare the actual measured values ​​and predicted values ​​of the desulfurization system in real time, and drive the prediction model to achieve parameter self-update, thereby forming a complete intelligent optimization closed loop.

9. The system according to claim 8, characterized in that, The intelligent decision-making unit also includes an economic optimization algorithm module. The economic optimization algorithm module takes the ratio of potassium ion concentration to sodium ion concentration in the desulfurization system as input and automatically calculates the optimal solution replacement timing by comparing the long-term expected total cost under different decisions.

10. The system according to claim 9, characterized in that, The economic optimization algorithm module calculates the optimal replacement timing based on the dynamic programming principle, and its state transition equation is: in, V(t) This represents the minimum expected total cost starting from time t. The immediate cost of choosing to continue running at time t. The immediate cost of choosing to perform solution displacement at time t. Let represent the minimum expected total cost starting from the next decision time t+1. This represents the minimum expected total cost when the system returns to its initial state after solution displacement.

Images