A method and system for optimizing control of main reheat steam temperature of a thermal power unit

Abstract

This invention provides an optimized control method and system for the main reheat steam temperature of a thermal power unit, relating to the field of thermal power generation technology. The method includes the following steps: Step S1, real-time acquisition and preprocessing of operating parameters; Step S2, determination of the unit's current operating condition based on the preprocessed operating parameters, and matching the corresponding predictive control model; Step S3, inputting the preprocessed operating parameters into the predictive control model and outputting the predicted steam temperature value; Step S4, generating control commands based on the predicted steam temperature value, the current steam temperature setpoint, and the optimization objective function; Step S5, superimposing the control commands with PID feedback control quantities to generate the final execution command and sending it to the DCS system for execution; Step S6, monitoring the actual steam temperature, processing prediction errors, and updating the predictive control model. This invention, employing the above-mentioned optimized control method and system for the main reheat steam temperature of a thermal power unit, improves steam temperature control accuracy and enhances the system's response speed and robustness.

Description

Technical Field

[0001] This invention relates to the field of thermal power generation technology, and in particular to an optimized control method and system for the main reheat steam temperature of a thermal power unit. Background Technology

[0002] Currently, stable control of main steam temperature and reheat steam temperature is crucial for ensuring the safe and economical operation of large thermal power units. With the increasing demands from the power grid for flexible unit operation, units frequently face complex operating conditions such as wide load regulation, rapid load changes, pulverizer start-up and shutdown, and soot blowing, placing higher demands on the response speed, control accuracy, and robustness of the steam temperature control system. Traditional steam temperature control often employs a proportional-integral-derivative (PID) cascade control strategy, combined with simple static feedforward adjustment, which can still meet basic control requirements under steady-state conditions.

[0003] However, traditional control strategies face significant technical bottlenecks when dealing with complex dynamic conditions. First, steam temperature is inherently characterized by large time lag, strong coupling, and numerous external disturbances. The effect of desuperheating water regulation on steam temperature has a significant time lag, making it difficult for traditional PID control to respond in advance, easily leading to overshoot or undershoot. Second, changes on the boiler combustion side (such as fluctuations in fuel calorific value, air volume adjustments, coal mill combination switching, and soot blowing operations) significantly affect the balance of radiative and convective heat transfer within the furnace, thus disturbing the steam temperature. However, traditional control methods lack effective feedforward compensation for these disturbances. Furthermore, the dynamic characteristics of the unit differ significantly under different load conditions, making it difficult for a fixed-parameter control model to maintain excellent control quality across the entire load range. In addition, existing systems have limited utilization of operational data, lacking predictive and adaptive capabilities based on big data and artificial intelligence, making it difficult to achieve online optimization and parameter self-tuning of the control model.

[0004] Therefore, there is an urgent need to develop a main reheat steam temperature optimization control method and system that can adapt to complex dynamic operating conditions, overcome the effects of large lag and strong coupling, and achieve adaptive adjustment across the entire load range, so as to improve the flexibility and safety of unit operation while ensuring steam temperature control accuracy. Summary of the Invention

[0005] The purpose of this invention is to provide an optimized control method and system for the main reheat steam temperature of thermal power units, in order to solve the technical problems in the prior art caused by the inherent large lag, strong coupling, and multiple external disturbances of the steam temperature object, as well as the response lag, large overshoot, and low control accuracy of traditional PID cascade control under complex dynamic conditions. The invention achieves automatic identification and switching under different operating conditions, ensures the stability and consistency of steam temperature control quality across the entire load range, improves steam temperature control accuracy, and enhances the system's response speed and robustness.

[0006] To achieve the above objectives, the present invention provides an optimized control method for the main reheat steam temperature of a thermal power unit, comprising the following steps: Step S1: Real-time acquisition and preprocessing of operating parameters; Step S2: Determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model; Step S3: Input the preprocessed operating parameters into the predictive control model and output the predicted steam temperature value; Step S4: Generate control commands based on the predicted steam temperature, the current steam temperature setpoint, and the optimized objective function; Step S5: Superimpose the control command with the PID feedback control quantity to generate the final execution command and send it to the DCS system for execution; Step S6: Monitor the actual steam temperature, process the prediction error, and update the predictive control model.

[0007] Preferably, step S6 further includes returning to step S3 based on the updated predictive control model in the next sampling period to achieve rolling optimization control.

[0008] Preferably, in step S2, the predictive control model is constructed using a combination of mechanistic modeling and data-driven modeling. Mechanistic modeling establishes a mathematical model framework based on heat transfer and thermodynamic principles, encompassing in-furnace radiative heat transfer, convective heat transfer, and steam flow processes, to describe the dynamic relationship between steam temperature, flow rate, and control variables. Data-driven modeling uses a long short-term memory network to identify and optimize the parameters of the mathematical model framework, and trains the model using pre-collected historical operating parameters. During training, the pre-collected historical operating parameters are divided into training, validation, and test sets, and a stochastic gradient descent optimization algorithm is used, with mean squared error as the loss function, to iteratively train and adjust the model parameters on the training set.

[0009] Preferably, in step S4, the objective function is a multi-objective optimization function; when constructing the multi-objective optimization function, a multi-objective optimization strategy is adopted, which comprehensively considers minimizing steam temperature deviation, minimizing fuel consumption, and maximizing equipment life; a particle swarm optimization algorithm or a non-dominated sorting genetic algorithm is used to search for the optimal control strategy that satisfies multiple objectives, which serves as the control command.

[0010] Preferably, in step S4, the control command includes a desuperheating water regulating valve command; a wall temperature limit is introduced in the desuperheating water regulation, and the opening range of the desuperheating water regulating valve is predicted through big data analysis to avoid over-adjustment and under-adjustment.

[0011] Preferably, in step S4, the control command also includes the flue gas damper opening adjustment amount, which automatically adjusts the flue gas damper opening based on the predicted steam temperature and the actual control requirements of the main reheat steam temperature; at the same time, it monitors the combustion parameters and coordinates the adjustment of the burner parameters through the combustion optimization control algorithm to maintain stable combustion.

[0012] Preferably, in step S6, a Kalman filter is used to process the prediction error between the actual steam temperature and the predicted steam temperature; and an online parameter identification algorithm is used to update the parameters of the predictive control model.

[0013] This invention also provides an optimized control system for the main reheat steam temperature of a thermal power unit, employing the above-mentioned method, including: The data interface module is used to communicate with the DCS system via the OPC protocol, collect operating parameters, and issue control commands. A real-time database is used to store and manage the operational parameters collected in real time. The operating condition identification module is used to determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model. Predictive control model is used to process preprocessed operating parameters and output predicted steam temperature values; An optimization solver is used to solve for control commands based on an optimization objective function. The collaborative control module is used to superimpose control commands with PID feedback control quantities to generate the final execution command; The parameter adaptation module is used to update the parameters of the predictive control model based on the online parameter identification algorithm; The safety protection module is used to monitor abnormal system states and seamlessly switch back to DCS system control in the event of an abnormality.

[0014] Preferably, the security protection module includes: The communication verification unit is used to automatically cut off the optimized control and seamlessly switch back to DCS system control in the event of a communication failure. The fault diagnosis unit is used to automatically disconnect the corresponding circuit of the system when a fault is detected in the signal measurement point or the execution variable, and to automatically disconnect the system and alarm when there is an MFT action, equipment failure or a large change in load. The command limiting unit is used to limit the upper and lower limits of the adjustment amount of the control command and the rate limiting protection, and to control whether the control command is output to the actuator. The commissioning and disconnection unit is used to perform commissioning and disconnection operations on the optimized control system on the DCS system, and to achieve disturbance-free tracking during system switching.

[0015] Therefore, the present invention employs the above-mentioned optimized control method and system for the main reheat steam temperature of thermal power units, and the beneficial technical effects are as follows: (1) By adopting a collaborative strategy that superimposes control commands and PID feedback control quantities, this invention effectively overcomes the large lag characteristics of the steam temperature object and greatly improves the response speed and control accuracy of the desuperheating water regulation.

[0016] (2) This invention constructs a predictive control model that combines mechanistic modeling and data-driven modeling. Mechanistic modeling establishes a mathematical model framework that includes in-furnace radiation heat transfer, convection heat transfer, and steam flow processes. Data-driven modeling uses a long short-term memory network to identify and optimize model parameters, enabling the model to accurately reflect the dynamic characteristics of the system and achieve advanced prediction of steam temperature changes in future periods, thereby improving the quality of steam temperature control under dynamic operating conditions.

[0017] (3) This invention designs an operating condition identification module to monitor operating parameters in real time, accurately determine the current operating condition of the unit and match the corresponding predictive control model. When the operating condition of the unit changes, it can immediately trigger the control switching strategy to ensure the stability and consistency of steam temperature control quality in the full load range.

[0018] (4) The present invention uses an online parameter identification algorithm to update the parameters of the predictive control model in real time. When the unit load changes significantly, the model parameters can be quickly adjusted so that the control model can better adapt to the new operating conditions. At the same time, when the operating conditions change significantly, it can automatically switch to a pre-trained predictive control model that is suitable for the operating conditions, which further improves the accuracy and stability of the control.

[0019] (5) The present invention introduces wall temperature limitation in desuperheating water regulation, and ensures the safety of equipment such as superheater and reheater under high temperature pressure by limiting the wall temperature.

[0020] (6) In the flue gas damper control, the present invention realizes the adjustment mechanism of automatically adjusting the opening of the flue gas damper according to the predicted value of steam temperature, while monitoring the combustion parameters, coordinating and adjusting the burner parameters through the combustion optimization control algorithm, maintaining combustion stability, and ensuring efficient combustion in the furnace while adjusting the steam temperature.

[0021] (7) By designing a safety protection module, the present invention monitors abnormal system states and seamlessly switches back to DCS system control in case of an abnormality, thus effectively ensuring the safe operation of the unit.

[0022] (8) This invention uses a Kalman filter to process the prediction error between the actual steam temperature and the predicted steam temperature, and uses an online parameter identification algorithm to update the parameters of the predictive control model, thereby realizing real-time estimation and compensation of model uncertainty and disturbance, continuously improving the accuracy of model prediction, and forming a rolling optimization control closed loop. Attached Figure Description

[0023] Figure 1This is a flowchart of an optimized control method for the main reheat steam temperature of a thermal power unit according to the present invention; Figure 2 This is a data architecture diagram of an optimized control system for the main reheat steam temperature of a thermal power unit according to the present invention. Figure 3 Flowchart of intelligent optimization control for main steam temperature; Figure 4 A schematic diagram of a primary desuperheating water intelligent control model; Figure 5 This is a schematic diagram of a two-stage desuperheating water intelligent control model; Figure 6 Trend chart of main reheat steam temperature optimization control effect under wide load conditions; Figure 7 Trend chart showing the effect of optimized control of main reheat steam temperature under stable load conditions. Detailed Implementation

[0024] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0025] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0026] Example 1 like Figure 1 As shown, an optimized control method for the main reheat steam temperature of a thermal power unit includes the following steps: Step S1: Collect and preprocess operating parameters in real time.

[0027] During unit operation, operating parameters, including steam flow, temperature, pressure, furnace temperature distribution, and flue gas composition, are first collected in real time through sensors and a DCS (Distributed Control System). The collected operating parameters are then transmitted to a real-time database for storage and management via the OPC protocol.

[0028] In the preprocessing stage, the collected operating parameters are cleaned to remove outliers and noise interference, and then standardized through normalization and other methods to obtain the preprocessed operating parameters.

[0029] Furthermore, based on domain knowledge and data analysis, features closely related to radiative and convective heat transfer within the furnace, such as the rate of change of steam enthalpy and the temperature gradient of flue gas in different regions, are extracted from the preprocessed operating parameters. These features, as inputs to the predictive control model, can more accurately reflect dynamic changes within the furnace and improve the model's prediction accuracy.

[0030] Step S2: Determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model.

[0031] Based on the preprocessed operating parameters, the current operating condition of the unit is determined through pattern recognition algorithms or expert systems. Specific operating conditions include low load, medium load, high load, start-up, shutdown, and rapid load change.

[0032] Based on the operating condition identification results, a predictive control model matching the current operating condition is invoked. The predictive control model is pre-trained for different operating conditions to ensure optimal control performance under various operating conditions.

[0033] When significant changes occur in the unit's operating conditions (such as switching from low-load to high-load conditions), an automatic model switching strategy is triggered to switch to a predictive control model suitable for the new operating condition. For example, when the unit switches from a stable operating state to a rapid load change state, if the load change rate exceeds a set threshold, the control switching strategy is immediately triggered, switching the control strategy from a steady-state control strategy to a load change control strategy, and adjusting the predictive control model parameters accordingly to ensure stable steam temperature control during the load change process.

[0034] Specifically, the predictive control model is constructed using a combination of mechanistic modeling and data-driven modeling. Mechanistic modeling establishes a mathematical framework based on heat transfer and thermodynamic principles, encompassing in-furnace radiative heat transfer, convective heat transfer, and steam flow processes, to describe the dynamic relationship between steam temperature, flow rate, and control variables. Data-driven modeling employs a Long Short-Term Memory (LSTM) network to identify and optimize the parameters of the mathematical model framework, and trains the model using pre-collected historical operating parameters. During training, the pre-collected historical operating parameters are divided into training, validation, and test sets, and a stochastic gradient descent optimization algorithm is used, with mean squared error as the loss function, for iterative training on the training set to adjust the model parameters. Furthermore, during training, visualization tools are used to monitor metrics such as the loss function and accuracy in real time to ensure model convergence and avoid overfitting.

[0035] During real-time computation and analysis, intermediate results can be stored in a key-value database for quick access, while structured and historical data can be stored in a relational database for subsequent analysis and querying. This multi-level storage architecture ensures efficient data access and the system's real-time responsiveness.

[0036] Step S3: Input the preprocessed operating parameters into the predictive control model and output the predicted steam temperature value.

[0037] After the model training is completed, the preprocessed operating parameters are received in real time, and the predicted steam temperature values ​​for multiple future sampling periods are output.

[0038] Step S4: Generate control commands based on the predicted steam temperature, the current steam temperature setpoint, and the optimized objective function.

[0039] The objective function is optimized as a multi-objective optimization function. When constructing the multi-objective optimization function, a multi-objective optimization strategy is adopted, which comprehensively considers minimizing steam temperature deviation, minimizing fuel consumption, and maximizing equipment life. Particle swarm optimization algorithm or non-dominated sorting genetic algorithm is used to search for the optimal control strategy that satisfies multiple objectives, which serves as the control command.

[0040] The generated control commands include commands for the desuperheating water regulating valve, the adjustment amount of the flue gas damper opening, and the adjustment amount of the fuel quantity.

[0041] The desuperheating water regulating valve command is used to control the opening degree of the desuperheating water regulating valve, including primary desuperheating water control and secondary desuperheating water control. For example... Figure 3 As shown, by combining expert experience and big data analysis, the main steam temperature intelligent optimization control model algorithm generates desuperheating water regulating valve commands, enabling the desuperheating water regulating valve to act in advance before the steam temperature changes, effectively overcoming the large lag characteristic of desuperheating water control and achieving rapid and accurate regulation of desuperheating water.

[0042] Specifically, in the primary cooling water control, such as Figure 4 As shown, based on the deviation between the intermediate point temperature setpoint and the intermediate point temperature measurement, and combined with combustion factors (such as total coal feed and total air volume), the first-level desuperheating water intelligent control model algorithm calculates the first-level desuperheating water regulating valve command to achieve precise control.

[0043] In secondary desuperheating water control, such as Figure 5 As shown, based on the deviation between the current steam temperature setpoint and the measured main steam temperature, and combined with combustion factors, the secondary desuperheating water intelligent control model algorithm calculates the secondary desuperheating water regulating valve command to further finely adjust the main steam temperature.

[0044] Introducing wall temperature limits into the desuperheating water regulation process and using big data analysis to predict the opening range of the desuperheating water regulating valve can prevent over-adjustment and under-adjustment.

[0045] The flue gas damper opening adjustment is used to regulate the balance between radiative and convective heat transfer. Based on the predicted steam temperature and the actual control requirements of the main and reheat steam temperatures, the flue gas damper opening is automatically adjusted. Specifically, when it is predicted that enhanced radiative heat transfer within the furnace may lead to an increase in the main and reheat steam temperature, the flue gas damper is opened wider to increase the flue gas flow in the convective heat transfer zone, allowing more heat to be transferred to the steam via convection, thus balancing the impact of radiative heat transfer and stabilizing the main and reheat steam temperature. Conversely, when it is predicted that excessive convective heat transfer may result in an excessively high main and reheat steam temperature, the flue gas damper is closed wider to reduce convective heat transfer and ensure that the main and reheat steam temperature remains within a reasonable range.

[0046] Meanwhile, adjusting the flue gas damper opening affects the flue gas flow and combustion conditions within the furnace, thus requiring coordinated control with the combustion system. When adjusting the flue gas damper opening, combustion parameters such as furnace negative pressure and oxygen content are monitored simultaneously. Through a combustion optimization control algorithm, parameters such as burner air distribution and fuel quantity are adjusted accordingly to ensure stable and efficient combustion within the furnace while regulating steam temperature. For example, when opening the flue gas damper too wide causes a decrease in furnace negative pressure, the air supply is automatically increased to maintain stable furnace negative pressure. Simultaneously, the fuel quantity is adjusted appropriately based on changes in oxygen content to ensure complete combustion and avoid adverse effects on the combustion system caused by flue gas damper adjustments.

[0047] Fuel quantity adjustment is used to coordinate the control of the combustion system and maintain a stable intermediate point temperature. For intermediate point temperature control, fuzzy control rules and predictive control methods are combined to adjust the fuel quantity and feedwater flow rate by monitoring operating parameters such as intermediate point temperature, steam flow rate, and load, thereby maintaining the intermediate point temperature within the set range.

[0048] Specifically, fuzzy control rules are used to process the intermediate point temperature deviation and its rate of change to obtain a preliminary direction and magnitude for fuel quantity adjustment. For example, when the intermediate point temperature is higher than its set value and the rate of change of deviation is large, the fuzzy control rules suggest that the fuel quantity should be appropriately reduced. Then, combined with a predictive control model, the trend of intermediate point temperature change under this fuel quantity adjustment is predicted based on the unit's dynamic characteristics and current operating conditions. If the prediction result shows that the intermediate point temperature can stabilize near the set value, the fuel quantity adjustment is executed; if the prediction result is not ideal, the fuel quantity adjustment output by the fuzzy control rules is corrected, and the prediction is repeated until the optimal control scheme is obtained.

[0049] Meanwhile, the intermediate point temperature is closely related to the feedwater flow rate, thus requiring coordinated control with the feedwater system. Based on the control requirements of the intermediate point temperature, the corresponding feedwater flow rate adjustment is calculated using a predictive control model. This adjustment also considers the influence of factors such as steam flow rate and load, enabling precise regulation of the feedwater flow rate. For example, when the intermediate point temperature rises and a suitable increase in feedwater flow rate is needed, the predictive control model calculates a feedwater flow rate adjustment value that both lowers the intermediate point temperature and ensures that steam production meets load requirements, based on the current steam flow rate and load conditions. This value is then sent to the feedwater control system for execution to compensate for the impact of fuel quantity changes.

[0050] Based on the predicted steam temperature, the setpoints of control variables such as desuperheating water flow rate and flue gas damper opening are adjusted in advance to cope with upcoming load changes and operating condition fluctuations. For example, if the predicted steam temperature will rise by 3 degrees Celsius within the next 3 minutes... In this case, the flow rate of the desuperheating water is gradually increased in advance, while the opening of the flue gas damper is finely adjusted to suppress the rising trend of steam temperature.

[0051] Step S5: Superimpose the control command with the PID feedback control quantity to generate the final execution command and send it to the DCS system for execution.

[0052] The control command is used as a feedforward signal, and a PID controller is used to provide feedback control on the deviation between the actual steam temperature and the current steam temperature setpoint, generating a PID feedback control quantity. The control command and the PID feedback control quantity are then superimposed to generate the final execution command.

[0053] For example, when it is predicted that the main reheat steam temperature will rise in the near future, the required increase in desuperheating water flow is calculated in advance based on the predicted steam temperature value. This calculated flow rate is the control command and is sent as a feedforward signal. Simultaneously, the deviation between the actual steam temperature and the current setpoint is monitored in real time. The PID controller generates a PID feedback control quantity based on the magnitude and rate of change of the deviation. The control command and the PID feedback control quantity are then superimposed to generate the final execution command.

[0054] If the control command increases the desuperheating water flow rate by a certain amount, but the actual steam temperature is still higher than the current steam temperature setpoint, the PID controller will further increase the PID feedback control quantity, so that the superimposed final execution command further increases the desuperheating water flow rate, thereby quickly suppressing the rising trend of steam temperature and making the steam temperature return to near the current steam temperature setpoint as soon as possible.

[0055] The generated final execution command is transmitted back to the real-time database via the OPC protocol, and then sent to the DCS system, which controls the actions of actuators such as the desuperheating water regulating valve and the flue gas damper.

[0056] Step S6: Monitor the actual steam temperature, process the prediction error, and update the predictive control model.

[0057] The actual steam temperature after execution is monitored in real time and compared with the predicted steam temperature value to calculate the prediction error. Then, a Kalman filter is used to process the prediction error, obtaining an estimate of model uncertainty and external disturbances. Based on this estimate, the predictive control model is corrected online to compensate for model errors and improve the prediction accuracy for the current cycle.

[0058] Simultaneously, online parameter identification algorithms are employed to update the parameters of the predictive control model, improving the accuracy of model predictions. Specifically, algorithms such as recursive least squares or extended Kalman filtering are used to continuously update model parameters based on real-time acquired operating parameters, enabling the model to adapt to changes in unit operating conditions. To improve the accuracy and stability of parameter identification, a forgetting factor is used to weight historical operating parameters, giving currently acquired operating parameters a greater influence on model parameter updates, thus enabling the model to quickly adapt to changes in unit operating conditions.

[0059] The system monitors changes in unit operating parameters in real time. When significant changes in operating parameters, such as load changes exceeding a certain threshold or sudden changes in steam flow, are detected, the model parameter adaptive adjustment process is triggered. The updated model parameters are applied to the control algorithm, the control input is recalculated, and the performance of the online parameter identification algorithm is evaluated by monitoring the control effect (such as steam temperature deviation and the smoothness of control input changes). If the adjusted control effect is found to be unsatisfactory, the relevant parameters of the online parameter identification algorithm (such as the forgetting factor) are automatically adjusted, and the online parameter identification process is re-executed until the ideal control effect is obtained.

[0060] After the model is updated, in the next sampling period, the updated predictive control model is used to return to step S3 for re-prediction and optimization, thus achieving rolling optimization control.

[0061] An optimized control system for the main reheat steam temperature of a thermal power unit, employing the aforementioned optimized control method for the main reheat steam temperature of a thermal power unit, has the following data architecture: Figure 2 As shown, the system communicates with the DCS system via the OPC protocol. Collected operating parameters are stored in a real-time database; intermediate calculation results are stored in a KV database for quick access; and structured and historical data are stored in a relational database for easy subsequent analysis and querying. The ICS intelligent control platform performs real-time calculations and analysis based on big data and AI algorithms, enabling wide-load cruise and control optimization. The generated control commands are sent back to the DCS system for execution via the OPC protocol.

[0062] An optimized control system for the main reheat steam temperature of a thermal power unit includes the following modules: The data interface module is used to communicate with the DCS system via the OPC protocol, collect operating parameters, and issue control commands.

[0063] A real-time database is used to store and manage the runtime parameters collected in real time.

[0064] The operating condition identification module is used to determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model.

[0065] A predictive control model is used to process the preprocessed operating parameters and output predicted steam temperature values.

[0066] An optimization solver is used to solve for control commands based on an optimization objective function.

[0067] The collaborative control module is used to superimpose control commands with PID feedback control quantities to generate the final execution command.

[0068] The parameter adaptation module is used to update the parameters of the predictive control model based on the online parameter identification algorithm.

[0069] The safety protection module is used to monitor abnormal system states and seamlessly switch back to DCS system control in the event of an anomaly, including: The communication verification unit is used to automatically disconnect the optimized control and seamlessly switch back to DCS system control in the event of a communication failure.

[0070] The fault diagnosis unit is used to automatically disconnect the corresponding circuit of the system when a fault is detected in the signal measuring point (such as steam temperature sensor, steam pressure sensor, flow meter, etc.) or the executed variable (such as desuperheating water regulating valve command, flue gas damper opening adjustment amount, etc.), and to automatically disconnect the system and alarm when MFT action, equipment failure, or large load change.

[0071] The command limiting unit is used to limit the upper and lower limits of the adjustment amount of the control command and the rate limiting protection, and to control whether the control command is output to the actuator.

[0072] The commissioning and disconnection unit is used to perform commissioning and disconnection operations on the optimized control system on the DCS system, and to achieve disturbance-free tracking during system switching.

[0073] All abnormalities, faults, and switching actions will be alarmed or prompted in the DCS system's operation screen, allowing operators to monitor the system status in real time.

[0074] To verify the control effect of this invention in an actual unit, Unit 7 of a thermal power plant was used as the implementation object. Field tests were conducted under both wide-load variable operating conditions and stable load operating conditions. The test results are as follows: Figure 6 , Figure 7 As shown.

[0075] like Figure 5 As shown, this invention remained stably operational during a wide load-variable operating condition period, with the unit's power generation gradually increasing from approximately 290MW to approximately 550MW. Test results indicate that the actual main reheat steam temperature consistently hovered around... The setpoint fluctuates smoothly, and the steam temperature deviation is less than [value missing] under steady-state conditions. Under dynamic operating conditions such as rapid load changes and load adjustments, the steam temperature deviation is less than [a certain value]. The main reheat desuperheating water regulating valve automatically and precisely adjusts according to load, combustion and steam temperature changes. The system automatically starts operation across the entire load range, effectively overcoming the impact of large load changes and combustion disturbances on reheat steam temperature. It also solves the technical defects of traditional control schemes, such as large overshoot, slow response and large fluctuation, under variable load conditions.

[0076] like Figure 6 As shown, under stable operating conditions with the unit's power generation power remaining between 298MW and 303MW and slight fluctuations in furnace oxygen levels, this invention was continuously put into operation. Test results show that the average main reheat steam temperature is approximately... ,and The setpoints are highly compatible, and the steady-state steam temperature deviation is controlled within [specific range]. Within a certain range, the steady-state fluctuation performance is far superior to that of conventional control schemes. Under conditions of minor disturbances in oxygen content and combustion status, the main reheat steam temperature shows no significant fluctuations, and the desuperheating water regulating valve outputs stable commands, achieving high-precision, overshoot-free stable control of the main reheat steam temperature under stable operating conditions.

[0077] In summary, this invention can improve the control accuracy and stability of the main reheat steam temperature, reduce steam temperature fluctuations under steady-state and dynamic operating conditions, meet the requirements of wide-load cruise and flexible operation of the unit, and at the same time reduce desuperheating water consumption, thereby improving the unit's operating economy and equipment safety.

[0078] Therefore, the present invention adopts the above-mentioned optimized control method and system for the main reheat steam temperature of thermal power units, which solves the technical problems in the prior art caused by the inherent large lag, strong coupling, and multiple external disturbances of the steam temperature object, as well as the response lag, large overshoot, and low control accuracy of traditional PID cascade control under complex dynamic conditions. It realizes automatic identification and switching under different operating conditions, ensures the stability and consistency of steam temperature control quality across the entire load range, improves steam temperature control accuracy, and enhances the system's response speed and robustness.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An optimized control method for the main reheat steam temperature of a thermal power unit, characterized in that, Includes the following steps: Step S1: Real-time acquisition and preprocessing of operating parameters; Step S2: Determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model; Step S3: Input the preprocessed operating parameters into the predictive control model and output the predicted steam temperature value; Step S4: Generate control commands based on the predicted steam temperature, the current steam temperature setpoint, and the optimized objective function; Step S5: Superimpose the control command with the PID feedback control quantity to generate the final execution command and send it to the DCS system for execution; Step S6: Monitor the actual steam temperature, process the prediction error, and update the predictive control model.

2. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 1, characterized in that, Step S6 also includes returning to step S3 based on the updated predictive control model in the next sampling period to achieve rolling optimization control.

3. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 1, characterized in that, In step S2, the predictive control model is constructed using a combination of mechanistic modeling and data-driven modeling. Mechanistic modeling establishes a mathematical model framework based on heat transfer and thermodynamic principles, encompassing in-furnace radiative heat transfer, convective heat transfer, and steam flow processes, to describe the dynamic relationship between steam temperature, flow rate, and control variables. Data-driven modeling uses a long short-term memory network to identify and optimize the parameters of the mathematical model framework, and trains the model using pre-collected historical operating parameters. During training, the pre-collected historical operating parameters are divided into training, validation, and test sets, and a stochastic gradient descent optimization algorithm is used, with mean squared error as the loss function, to iteratively train and adjust the model parameters on the training set.

4. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 1, characterized in that, In step S4, the objective function is a multi-objective optimization function. When constructing the multi-objective optimization function, a multi-objective optimization strategy is adopted, which comprehensively considers minimizing steam temperature deviation, minimizing fuel consumption, and maximizing equipment life. The particle swarm optimization algorithm or the non-dominated sorting genetic algorithm is used to search for the optimal control strategy that satisfies multiple objectives, which serves as the control command.

5. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 4, characterized in that, In step S4, the control commands include commands for the desuperheating water regulating valve; a wall temperature limit is introduced in the desuperheating water regulation, and the opening range of the desuperheating water regulating valve is predicted through big data analysis to avoid over-adjustment and under-adjustment.

6. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 4, characterized in that, In step S4, the control command also includes the adjustment amount of the flue gas damper opening. Based on the predicted steam temperature and the actual control requirements of the main reheat steam temperature, the flue gas damper opening is automatically adjusted. At the same time, combustion parameters are monitored, and the burner parameters are coordinated and adjusted through the combustion optimization control algorithm to maintain combustion stability.

7. The optimized control method for main reheat steam temperature of a thermal power unit according to claim 1, characterized in that, In step S6, a Kalman filter is used to process the prediction error between the actual steam temperature and the predicted steam temperature; an online parameter identification algorithm is used to update the parameters of the predictive control model.

8. An optimized control system for the main reheat steam temperature of a thermal power unit, comprising the method described in any one of claims 1-7, characterized in that, include: The data interface module is used to communicate with the DCS system via the OPC protocol, collect operating parameters, and issue control commands. A real-time database is used to store and manage the operational parameters collected in real time. The operating condition identification module is used to determine the current operating condition of the unit based on the preprocessed operating parameters and match the corresponding predictive control model. Predictive control model is used to process preprocessed operating parameters and output predicted steam temperature values; An optimization solver is used to solve for control commands based on an optimization objective function. The collaborative control module is used to superimpose control commands with PID feedback control quantities to generate the final execution command; The parameter adaptation module is used to update the parameters of the predictive control model based on the online parameter identification algorithm; The safety protection module is used to monitor abnormal system states and seamlessly switch back to DCS system control in the event of an abnormality.

9. An optimized control system for the main reheat steam temperature of a thermal power unit according to claim 8, characterized in that, The security protection module includes: The communication verification unit is used to automatically disconnect the optimized control and seamlessly switch back to DCS system control in the event of a communication failure. The fault diagnosis unit is used to automatically disconnect the corresponding circuit of the system when a fault is detected in the signal measurement point or the execution variable, and to automatically disconnect the system and alarm when there is an MFT action, equipment failure or a large change in load. The command limiting unit is used to limit the upper and lower limits of the adjustment amount of the control command and the rate limiting protection, and to control whether the control command is output to the actuator. The commissioning and disconnection unit is used to perform commissioning and disconnection operations on the optimized control system on the DCS system, and to achieve disturbance-free tracking during system switching.

Images