Control method and system for double reheat unit to quickly respond to power grid demands

Abstract

The present invention pertains to the field of power grid balance control, and particularly relates to a control method and system for a double reheat unit to quickly respond to power grid demands. The method comprises: acquiring boiler-side and turbine-side operation data via integrated sensors, performing preprocessing and then constructing and training a multivariable predictive control model, and acquiring boiler-side initial control parameters; second, by introducing a dynamic compensation strategy, acquiring boiler-side real-time heat flow distribution, and optimizing the boiler-side initial control parameters to ensure the accuracy of boiler-side control; third, acquiring turbine-side initial control parameters by using a turbine-side peak shaving strategy and a particle swarm algorithm; and fourth, constructing a wind and photovoltaic power output prediction model, and dynamically adjusting the output of the double reheat unit on the basis of the prediction result, and when the required output exceeds a threshold, applying pre-calculated boiler-side and turbine-side control parameters to control the unit to perform power generation, so as to ensure a quick response to power grid demands. The present invention effectively improves the response speed and operational efficiency of the double reheat unit, satisfying the requirements of rapid power grid changes.

Description

Control methods and systems for rapid response of double reheat units to grid demand

[0001] This application claims priority to Chinese Patent Application No. 202411420493.4, filed on October 12, 2024, entitled "Control Method and System for Rapid Response to Grid Demand of Double Reheat Unit", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This invention belongs to the field of power grid balance control, and in particular relates to a control method and system for rapid response of secondary reheat units to power grid demand. Background Technology

[0003] With the transformation of the energy structure and the continuous increase in the proportion of renewable energy, the power grid faces unprecedented peak-shaving challenges. Double reheat ultra-supercritical coal-fired units, as the current technological frontier in combined heat and power (CHP), have become a key force supporting the stable operation of the power grid due to their high efficiency and low emissions. These units significantly improve the thermodynamic cycle efficiency through two steam reheat processes and demonstrate significant advantages in reducing pollutant emissions. However, despite the significant achievements of double reheat technology in efficiency improvement, it still faces some technical and operational limitations in responding to the rapidly changing demands of the power grid.

[0004] For example, the patent with authorization announcement number CN112910017B discloses a primary frequency regulation method for ultra-supercritical double reheat units under grid power shortage. First, the difference between the grid frequency and the standard frequency is calculated, and the opening degree of the primary reheater bypass valve is calculated according to the frequency difference. The primary reheater bypass valve is opened to a certain degree, and some steam bypasses the primary reheater and directly enters the high-pressure cylinder, which increases the working speed of steam in the turbine, thereby improving the primary frequency regulation response speed of the double reheat unit and enabling the grid frequency to return to normal value as soon as possible. In order to prevent the primary reheater from overheating, when the temperature of the primary reheater wall tube reaches the alarm value, the primary reheater bypass valve is fully closed by overshoot.

[0005] The existing technologies have the following problems: 1) The limited reheat steam temperature control on the furnace side and the peak-shaving capacity on the turbine side have become the main bottlenecks restricting the flexibility of the unit; 2) For the rapid response to wind and solar power balance during peak load periods, it is impossible to achieve intelligent control and realize accurate and rapid load tracking and automatic adjustment. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention proposes a control method for double reheat units to rapidly respond to grid demands. This method acquires boiler-side and turbine-side operating data through integrated sensors, preprocesses the data, and constructs and trains a multivariate predictive control model to achieve optimal boiler-side control. Secondly, it introduces a dynamic compensation strategy to optimize real-time heat flux distribution on the boiler side, ensuring the accuracy of boiler-side control. Thirdly, it utilizes a turbine-side peak-shaving strategy and particle swarm optimization algorithm to obtain optimal turbine-side control parameters. Fourthly, it constructs a wind and solar power output prediction model and dynamically adjusts the double reheat unit's output based on the prediction results. When the output demand exceeds a threshold, it applies pre-set boiler-side and turbine-side control parameters to control the unit for power generation, ensuring rapid response to grid demands. This invention effectively improves the response speed and operating efficiency of double reheat units, meeting the rapidly changing demands of the grid.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] Control methods for rapid response to grid demand in double reheat units include:

[0009] Step S1: Acquire boiler-side and turbine-side attribute data, monitoring operation data and load demand power data of the secondary reheat unit through integrated sensors, and preprocess the collected data;

[0010] Step S2: Construct a multivariate predictive control model. Based on the preprocessed furnace-side operating data and the multivariate predictive control model, obtain the initial control variable parameters of the furnace side.

[0011] Step S3: Introduce a dynamic compensation strategy to obtain the internal temperature field of the furnace. Based on the obtained internal temperature field of the furnace and the dynamic compensation strategy, obtain real-time heat flow distribution information on the furnace side. Feed the obtained real-time heat flow distribution information back to the multivariate predictive control model to compensate and correct the obtained initial control variable parameters.

[0012] Step S4: Introduce a turbine-side peak shaving strategy. Based on the turbine-side attribute data, use an evaluation algorithm to evaluate and obtain the initial peak shaving limit of the steam turbine. At the same time, based on the obtained initial peak shaving limit and turbine-side operating data, use a parameter search model constructed by the particle swarm optimization algorithm to obtain the initial control variable parameters of the turbine side.

[0013] Step S5: Construct and train the wind and solar power output prediction model. Based on the acquired wind and solar power output data and load demand power data, use the wind and solar power output prediction model to calculate the output power of the secondary reheat unit.

[0014] Step S6: Set the output threshold of the secondary reheat unit to 0, and feed back the obtained output power of the secondary reheat unit to the multivariate predictive control model and the turbine-side peak shaving strategy to obtain the furnace-side pre-control variable parameters and the turbine-side pre-control variable parameters.

[0015] Step S7: Compare the output power of the secondary reheat unit with the output threshold of the secondary reheat unit. If it is greater than the threshold, control the secondary reheat unit using the obtained optimal control variable parameters of the pre-heater side and the optimal control variable parameters of the pre-heater side. Generate electricity according to the calculated output power of the secondary reheat unit and output a power value equal to the output power of the secondary reheat unit to make up for the power missing from wind and solar power output. If it is less than or equal to the threshold, the secondary reheat unit remains in a stopped state.

[0016] Specifically, the specific steps for constructing the wind and solar power output prediction model in step S5 include:

[0017] S501. Collect historical wind and solar power output data, load balance power demand data, and weather factor data according to sunny, cloudy, overcast, and rainy weather conditions. Preprocess the collected data to construct a multi-weather-condition input sequence X. i =((x1…x n ), (y1…y m ), (z1…z k )); where X i Let x represent the data for the i-th weather condition. n y represents the nth wind and solar power output data in the i-th weather condition data. m z represents the m-th load balance power demand data in the i-th weather condition data. k This represents the k-th weather factor data in the i-th weather state data;

[0018] Specifically, the specific steps for constructing the wind and solar power output prediction model in step S5 also include:

[0019] S502. Using the random forest algorithm, a wind and solar power output prediction model is constructed in an integrated manner, including a sunny prediction sub-model, a cloudy prediction sub-model, an overcast prediction model, and a rain prediction sub-model.

[0020] S503, Input to multi-weather state input sequence X i =((x1…x n ), (y1…y m ), (z1…z k The weather conditions are input into the corresponding sub-models within the photovoltaic output prediction model for training.

[0021] S504. Using historical wind and solar power output data and corresponding load balance power demand data, obtain historical true output error; at the same time, using wind and solar power output data and load balance power demand data predicted at corresponding historical times, obtain historical predicted output error.

[0022] S504. Set the training threshold δ. Based on the historical actual output error and the historical predicted output error, obtain the secondary training error. Compare the obtained secondary training error with δ. If it is less than δ, the training is complete. Otherwise, continue training until the threshold condition is met.

[0023] Specifically, the calculation process for predicting the output power of the double reheat unit includes:

[0024] S505. Collect wind and solar power output data, load balance power demand data and weather factor data for the first 30 hours of the forecast day, and input the wind and solar power output data and weather factor data for the first 30 hours into the trained wind and solar power output forecast model, and output wind and solar power output data for the next 12 hours.

[0025] S506. Based on the wind and solar power output data and load balance power demand data for the next 12 hours, the output power of the secondary reheat unit is calculated using the load demand balance equation. The load demand balance equation is as follows: in, This represents the power required to balance the load at time t. This represents the predicted power output data of wind and solar power at time t. This represents the output power of the secondary reheat unit at time t.

[0026] Specifically, the specific steps of step S3 include:

[0027] S301. Construct a three-dimensional geometric model of the furnace side using furnace side attribute data, divide the three-dimensional geometric model of the furnace side into N monitoring areas using a mesh algorithm, and obtain the temperature distribution data of the N monitoring areas inside the furnace through deployed furnace side sensors.

[0028] S302. Based on the acquired temperature distribution data, the heat flow distribution map of each region in the furnace is calculated by combining finite element analysis with thermal radiation and hydrodynamic algorithms.

[0029] S303. Feed back the obtained heat flow distribution map of each region to the initial control variable parameters, use the dynamic temperature compensation formula to calculate the correction value of the initial control variable parameters, and compensate and correct the initial control variable parameters.

[0030] Specifically, the dynamic temperature compensation formula is as follows:

[0031] in, α represents the compensation amount for the i-th parameter of the initial control variable. t T represents the dynamic adjustment factor of the parameter. j T represents the real-time temperature data corresponding to the j-th monitoring area inside the furnace.rj K represents the expected temperature data corresponding to the j-th monitoring area inside the furnace, N represents the number of monitoring areas inside the furnace, and K represents the number of monitoring areas inside the furnace. p Indicates temperature deviation T j -T rj The corresponding correction parameter, K d Indicates the rate of temperature change The corresponding correction parameter, K i Indicates cumulative temperature deviation The corresponding correction parameter, t0, represents the length of time it takes for the secondary reheat turbine unit to operate once.

[0032] Specifically, the construction steps for a multivariate predictive control model include:

[0033] S201. Preprocess the furnace-side attribute data, operating data, and load demand power data acquired by the integrated sensors to obtain furnace-side control variable data;

[0034] S202. Construct a multivariate predictive control model using support vector machines, and input the acquired furnace-side control variable data into the multivariate predictive control model for training.

[0035] S203. Set the objective function and corresponding constraints of the multivariable predictive control model to limit the furnace side operating cost, operating efficiency and emissions, so as to minimize fuel consumption while maintaining the output power to meet the load power demand.

[0036] Specifically, the construction steps of a multivariate predictive control model also include:

[0037] S204. Integrate the constructed objective function and corresponding constraints into the multivariate predictive control model for training to obtain the trained multivariate predictive control model.

[0038] S205. Integrate the trained multivariate predictive control model into the furnace-side control subsystem, collect furnace-side operating data in real time through sensors, and input the real-time collected furnace-side operating data into the multivariate predictive control model to obtain the initial control variable parameters of the furnace side.

[0039] Specifically, the objective function and corresponding constraints in S203 include:

[0040] The specific formula for the objective function is as follows:

[0041] The constraints are:

[0042] Where J represents the furnace-side objective constraint function, E0 represents the fuel consumption cost for one furnace operation cycle, and E0 represents the total pollution emissions corresponding to one furnace operation cycle. c represents the actual output power of the double reheat unit at time t; t c represents the fuel consumption at time t. min c max f represents the maximum and minimum fuel consumption per unit time, respectively. t f represents the air velocity fed into the furnace at time t. min f max These represent the maximum and minimum values ​​of the air velocity fed into the furnace at time t, respectively. E represents the maximum power output threshold of the double reheat unit at time t. l The threshold represents the total amount of pollution emissions during one cycle of furnace operation, and w1, w2, and w3 represent the corresponding weighting coefficients.

[0043] Specifically, the formula for the initial peak-shaving interval corresponding to the preliminary peak-shaving limit in step S4 is as follows:

[0044] in, Indicates the initial peak-shaving interval, f a This represents the safety margin coefficient. This represents the minimum output power at which the steam turbine can operate stably at time t. This represents the additional energy loss caused by the turbine starting from a stopped state to full load or shutting down from full load at time t. This indicates the amount of output power that a steam turbine can increase or decrease per unit time.

[0045] Specifically, in step S4, the objective function and constraints for the turbine reaction are set in the parameter search model, as follows:

[0046] J r This represents the objective function for improving the machine-side operating efficiency. This represents the difference between the current output power of the secondary reheat unit and the previous output power of the secondary reheat unit. Indicates the corresponding time of change, e f This represents the efficiency of the steam turbine operation. This value is predicted using a steam turbine efficiency prediction model constructed with support vector machines based on the historical operating data of the steam turbine; w4 and w5 represent the corresponding weighting coefficients.

[0047] The control system for rapid response to grid demand of double reheat units includes: a data processing module, a boiler-side control module, a turbine-side control module, a wind and solar power output prediction module, and a feedback control module.

[0048] The data processing module is used to collect monitoring data and load demand power data from the furnace side and the machine side using integrated sensors, and to preprocess the collected data.

[0049] The wind and solar power output prediction module is used to build and train a wind and solar power output prediction model, calculate and predict wind and solar power output data based on the trained model, and calculate the output power of the secondary reheat unit by using the predicted wind and solar power output data and load demand power data.

[0050] Specifically, the furnace-side control module includes a multivariable control unit and a dynamic compensation unit;

[0051] The multivariable control unit is used to construct a multivariable predictive control model and obtain the initial control variable parameters of the furnace side based on the acquired furnace side operating data and control strategy. The dynamic compensation unit is used to construct a dynamic compensation strategy, calculate the real-time heat flow distribution information of the furnace side based on the acquired furnace internal temperature field information and compensation strategy, and use the heat flow distribution information to compensate and correct the initial control variable parameters of the furnace side.

[0052] Specifically, the machine-side control module includes a peak shaving assessment unit and a parameter calculation unit;

[0053] The peak shaving assessment unit is used to comprehensively assess the steam turbine based on the acquired turbine-side attribute data and an assessment algorithm to obtain the initial peak shaving limit of the steam turbine; the parameter calculation unit is used to obtain the initial control variable parameters of the turbine side based on the acquired initial peak shaving limit and turbine-side operating data through a particle swarm optimization algorithm.

[0054] Specifically, the feedback control module includes a parameter feedback optimization unit and a discrimination control unit;

[0055] The parameter feedback optimization unit is used to optimize the initial control variable parameters on the furnace side and the turbine side based on the obtained output power of the secondary reheat unit, so as to obtain the pre-control variable parameters on the furnace side and the turbine side. The discrimination control unit is used to set the output threshold of the secondary reheat unit and compare the threshold with the obtained output power of the secondary reheat unit to determine whether to generate electricity. If to generate electricity, the unit is controlled to generate electricity using the obtained pre-control variable parameters on the furnace side and the turbine side, and outputs a power value equal to the output power of the secondary reheat unit.

[0056] A computer-readable storage medium storing computer instructions that, when executed, provide a control method for a secondary reheat unit to rapidly respond to grid demands.

[0057] An electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a control method for a secondary reheat unit to quickly respond to grid demands.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] This invention addresses the shortcomings of existing technologies by constructing a multivariate predictive control model and introducing a dynamic compensation strategy. This enables more accurate prediction and regulation of the furnace-side operating parameters of the double reheat unit, particularly the furnace-side reheat steam temperature, thereby effectively overcoming the bottleneck of furnace-side reheat steam temperature control and improving the overall flexibility of the unit. Secondly, by introducing a turbine-side peak-shaving strategy and a parameter search model constructed using a particle swarm optimization algorithm, the peak-shaving limits of the turbine can be quickly assessed and optimized based on turbine-side attribute data and operating data, thereby enhancing the turbine-side peak-shaving capability and ensuring stable and efficient operation of the unit during peak load periods. Thirdly, through the construction and training of a wind and solar power output prediction model, accurate prediction of wind and solar power output can be achieved. Combined with load demand power data, the output power of the double reheat unit is calculated through the load demand balance equation. This process not only achieves rapid response to wind and solar power balance but also enables precise and rapid load tracking and automatic adjustment, ensuring that the unit maintains stable power output during fluctuations in wind and solar power output. Attached Figure Description

[0060] Figure 1 is a flowchart of the control method for rapid response of a double reheat unit to grid demand according to Embodiment 1 of the present invention.

[0061] Figure 2 is a control architecture diagram of the double reheat unit's rapid response to grid demand in Embodiment 1 of the present invention;

[0062] Figure 3 is a diagram of the wind and solar power output prediction model architecture of Embodiment 1 of the present invention;

[0063] Figure 4 is a control system module architecture diagram of the double reheat unit for rapid response to grid demand in Embodiment 2 of the present invention. Detailed Implementation

[0064] Example 1

[0065] Please refer to Figures 1 and 2. One embodiment of the present invention provides a control method for rapid response of a double reheat unit to grid demand, comprising the following steps:

[0066] Step S1: Acquire furnace-side and turbine-side attribute data, monitoring operation data, and load demand power data of the secondary reheat unit through integrated sensors, and preprocess the collected data; further, the furnace-side attribute data includes furnace-side geometry and structure, material properties, design parameters, and equipment specifications; furnace-side geometry and structure includes furnace dimensions (such as diameter, height, and volume), furnace lining material, number and arrangement of tuyeres, and cooling system structure (such as cooling walls, cooling plate positions and materials); material properties include the thermal conductivity, high-temperature resistance, and coefficient of thermal expansion of refractory materials; design parameters include rated capacity, Design fuel type, theoretical combustion air requirement, and design thermal efficiency; equipment specifications including the model, capacity, and performance parameters of auxiliary equipment such as fans, pumps, and valves; furnace side operating data including furnace internal temperature (temperature distribution at different monitoring points), furnace outlet temperature, furnace wall temperature, cooling water inlet and outlet temperatures, furnace pressure, furnace top pressure, furnace bottom pressure, flue pressure, steam pressure, fuel flow rate (pulverized coal, natural gas, etc.), combustion air flow rate, cooling water flow rate, thermal efficiency, actual output power, load rate, and combustion control parameters (such as air supply volume, fuel supply rate, and excess air coefficient);

[0067] Step S2: Construct a multivariate predictive control model. Based on the preprocessed furnace-side operating data and the multivariate predictive control model, obtain the initial control variable parameters of the furnace side.

[0068] Step S3: Introduce a dynamic compensation strategy to obtain the internal temperature field of the furnace. Based on the obtained internal temperature field of the furnace and the dynamic compensation strategy, obtain real-time heat flow distribution information on the furnace side. Feed the obtained real-time heat flow distribution information back to the multivariate predictive control model to compensate and correct the obtained initial control variable parameters.

[0069] Step S4: Introduce a turbine-side peak shaving strategy. Based on the turbine-side attribute data, use an evaluation algorithm to evaluate and obtain the initial peak shaving limit of the steam turbine. At the same time, based on the obtained initial peak shaving limit and turbine-side operating data, use a parameter search model constructed by the particle swarm optimization algorithm to obtain the initial control variable parameters of the turbine side.

[0070] Step S5: Construct and train the wind and solar power output prediction model. Based on the acquired wind and solar power output data and load demand power data, use the wind and solar power output prediction model to calculate the output power of the secondary reheat unit.

[0071] Step S6: Set the output threshold of the secondary reheat unit to 0, and feed back the obtained output power of the secondary reheat unit to the multivariate predictive control model and the turbine-side peak shaving strategy to obtain the furnace-side pre-control variable parameters and the turbine-side pre-control variable parameters.

[0072] Step S7: Compare the output power of the secondary reheat unit with the output threshold of the secondary reheat unit. If it is greater than the threshold, control the secondary reheat unit using the obtained optimal control variable parameters of the pre-heater side and the optimal control variable parameters of the pre-heater side. Generate electricity according to the calculated output power of the secondary reheat unit and output a power value equal to the output power of the secondary reheat unit to make up for the power missing from wind and solar power output. If it is less than or equal to the threshold, the secondary reheat unit remains in a stopped state.

[0073] Further, please refer to Figure 3. The specific steps for constructing the wind and solar power output prediction model in S5 include:

[0074] S501. Collect historical wind and solar power output data, load balance power demand data, and weather factor data according to sunny, cloudy, overcast, and rainy weather conditions. Preprocess the collected data to construct a multi-weather-condition input sequence X. i =((x1…x n ), (y1…y m ), (z1…z k )); where X i Let x represent the data for the i-th weather condition. n y represents the nth wind and solar power output data in the i-th weather condition data. m z represents the m-th load balance power demand data in the i-th weather condition data. k This represents the k-th weather factor data in the i-th weather state data; furthermore, in this embodiment, the lengths of the wind and solar power output data, load balance power demand data, and weather factor data in the multi-weather state input sequence are all 15 minutes;

[0075] S502. Using the random forest algorithm, a wind and solar power output prediction model is constructed in an integrated manner, including a sunny prediction sub-model, a cloudy prediction sub-model, an overcast prediction sub-model, and a rain prediction sub-model. Further, in this embodiment, the number of decision trees corresponding to the random forest algorithm in the sunny prediction sub-model, the cloudy prediction sub-model, the overcast prediction sub-model, and the rain prediction sub-model is set to 30.

[0076] S503, Input to multi-weather state input sequence X i =((x1…x n ), (y1…y m ), (z1…z k The weather conditions are input into the corresponding sub-models within the photovoltaic output prediction model for training.

[0077] S504. Using historical wind and solar power output data and corresponding load balance power demand data, obtain historical true output error; at the same time, using wind and solar power output data and load balance power demand data predicted at corresponding historical times, obtain historical predicted output error.

[0078] S504. Set a training threshold δ. Based on the historical actual output error and the historical predicted output error, obtain the secondary training error. Compare the obtained secondary training error with δ. If it is less than δ, the training is complete. Otherwise, continue training until the threshold condition is met. In this embodiment, the value of δ is 0.05.

[0079] S505. Collect wind and solar power output data, load balance power demand data and weather factor data for the first 30 hours of the forecast day, and input the wind and solar power output data and weather factor data for the first 30 hours into the trained wind and solar power output forecast model, and output wind and solar power output data for the next 12 hours.

[0080] S506. Based on the wind and solar power output data and load balance power demand data for the next 12 hours, the output power of the secondary reheat unit is calculated using the load demand balance equation. The load demand balance equation is as follows: in, This represents the power required to balance the load at time t. This represents the predicted power output data of wind and solar power at time t. This represents the output power of the secondary reheat unit at time t.

[0081] Furthermore, the specific process steps of S3 include:

[0082] S301. Construct a three-dimensional geometric model of the furnace side using furnace side attribute data. Divide the three-dimensional geometric model of the furnace side into N monitoring areas using a mesh algorithm, and acquire temperature distribution data of the N monitoring areas inside the furnace through deployed furnace side sensors. Further, this step is implemented in practice using 3D modeling software (such as SolidWorks, CATIA, etc.) and mesh generation software (such as GAMSCAD, NETGEN, ICEM, etc.). Specific steps include:

[0083] S3011. Input the detailed geometric information of the furnace side, including dimensions, shape, material, and structural layout data, into the 3D modeling software to construct a 3D geometric model of the furnace side.

[0084] S3012. Input the constructed three-dimensional geometric model of the furnace side into the mesh generation software, and set the mesh shape, mesh size and mesh density according to the fluid flow characteristics, heat conduction properties and furnace structure complexity. In this embodiment, a quadrilateral mesh shape is used to divide the constructed three-dimensional geometric model of the furnace side into a mesh.

[0085] S3013. Based on the grid division results, plan sensor deployment points at key locations in each monitoring area, collect temperature data through sensor networks or wired / wireless methods, and display the collected temperature data in the corresponding grid using different colors; for example, use red to mark high temperature areas, blue to mark low temperature areas, and transition areas to indicate transition colors, which are set by those skilled in the art according to usage habits.

[0086] S302. Based on the acquired temperature distribution data, the heat flow distribution map of each region in the furnace is calculated by combining finite element analysis with thermal radiation and hydrodynamic algorithms.

[0087] S303. Feedback the obtained heat flow distribution map of each region to the initial control variable parameters, calculate the correction value of the initial control variable parameters using the dynamic temperature compensation formula, and compensate and correct the initial control variable parameters. The specific dynamic temperature compensation formula is as follows:

[0088] in, α represents the compensation amount for the i-th parameter of the initial control variable. t T represents the dynamic adjustment factor of the parameter. j T represents the real-time temperature data corresponding to the j-th monitoring area inside the furnace. rj K represents the expected temperature data corresponding to the j-th monitoring area inside the furnace, N represents the number of monitoring areas inside the furnace, and K represents the number of monitoring areas inside the furnace. p Indicates temperature deviation T j -T rj The corresponding correction parameter, K d Indicates the rate of temperature change The corresponding correction parameter, K i Indicates cumulative temperature deviation The corresponding correction parameter, t0, represents the length of time for the secondary reheat turbine unit to operate once; further, in this embodiment... Incorporating the time rate of change enhances the control of dynamic response to temperature changes, particularly transient processes. By utilizing the cumulative effect of historical temperature deviations, long-term static errors can be eliminated, improving control accuracy. In this embodiment, parameter α... t K p K d and K i By utilizing historical furnace operation data, a genetic algorithm is used to periodically search for and obtain the parameter α. t K p K d and K i The corresponding optimal value; αt It can dynamically adjust the control strategy based on historical control performance, current operating condition stability, or the intensity of external disturbances, thus achieving the adaptability of the control strategy;

[0089] Furthermore, the specific steps for obtaining the heat flux distribution map in S302 of this embodiment include:

[0090] S3021. Based on the monitoring area defined in S301, a set of heat transfer equations in the furnace is constructed by combining the Navier-Stokes equation and the energy equation with the heat conduction, convection, and Rayleigh-Boltzmann equations. The technology involved in this equation is existing technology and will not be described in detail here. Those skilled in the art can set the set of equations based on the actual operating data of the furnace.

[0091] S3022. Based on the furnace-side attribute data and furnace-side operating requirements, set the furnace-side inlet boundary conditions, furnace wall boundary conditions, and furnace-side pressure boundary conditions. Further, in this embodiment, the furnace-side inlet boundary conditions, furnace wall boundary conditions, and pressure boundary conditions are basic attribute parameter constraints of the furnace side. These conditions are set by those skilled in the art based on the furnace-side attribute data, furnace-side operating requirements data, and the required output power data. Further, the boundary conditions in S3022 include:

[0092] Furnace side inlet boundary conditions: u t =u0, this condition specifies the fluid velocity entering the furnace, u t d represents the velocity of the fluid entering the furnace at time t, and u0 represents a fixed value set for the fluid velocity entering the furnace; t =d0, this condition is used to set the temperature of the inlet fluid, d t d0 represents the initial temperature at which the fluid enters the furnace at time t, and d0 represents the fixed initial temperature at which the fluid enters the furnace.

[0093] Furnace wall boundary conditions: This condition indicates that there is no heat exchange between the furnace wall and the outside environment, and the temperature gradient of the furnace wall is zero in the direction perpendicular to the surface. T tw This represents the surface temperature of the furnace wall at time t;

[0094] Furnace side pressure boundary conditions: This condition indicates that the rate of change of furnace pressure with respect to the boundary normal is a fixed value, and the furnace pressure changes uniformly, p t This indicates the pressure inside the furnace at time t;

[0095] S3023. Based on the equations and boundary conditions constructed above, solve the equations using finite element analysis software to calculate the heat flux distribution and temperature field, and obtain the heat flux density and direction; finite element analysis software includes: ANSYS, COMSOL, ABAQUS;

[0096] S3024. Based on the calculated heat flux distribution and temperature field, generate a heat flux distribution map using visualization software; the heat flux distribution map includes the furnace side temperature field, heat flow diagram, isotherms, and heat flux density diagram; visualization software includes: MATLAB, ParaView, and Tecplot;

[0097] Furthermore, the specific steps for obtaining the initial control variable parameters on the furnace side in step S2 of this embodiment include:

[0098] S201. Preprocess the furnace-side attribute data, operating data, and load demand power data acquired by the integrated sensors to obtain furnace-side control variable data;

[0099] S202. Construct a multivariate predictive control model using support vector machines, and input the acquired furnace-side control variable data into the multivariate predictive control model for training.

[0100] S203. Set the objective function and corresponding constraints of the multivariate predictive control model to limit furnace-side operating costs, operating efficiency, and emissions, so as to minimize fuel consumption while maintaining output power to meet load power demand. The specific formula of the objective function is: The constraints are:

[0101] Where J represents the furnace-side objective constraint function, E0 represents the fuel consumption cost for one furnace operation cycle, and E0 represents the total pollution emissions corresponding to one furnace operation cycle. c represents the actual output power of the double reheat unit at time t; t c represents the fuel consumption at time t. min c max f represents the maximum and minimum fuel consumption per unit time, respectively. t f represents the air velocity fed into the furnace at time t. min f max The values ​​represent the maximum and minimum wind velocities fed into the furnace at time t, respectively, where w1, w2, and w3 represent the corresponding weighting coefficients, and c... min c max and f min f max The data was obtained by calculating it using the random forest algorithm based on historical furnace operation data; E l The threshold represents the total amount of pollution emissions within one cycle of furnace operation. This threshold is set by the fixed amount of pollutants emitted per unit time as stipulated by the state. In this embodiment, the total amount of pollution emissions is the amount of carbon dioxide emitted by the furnace side per unit time multiplied by the length of one cycle of furnace operation. This represents the maximum power output threshold of the double reheat unit at time t. This threshold is determined and set based on the performance attributes of the double reheat unit and by those skilled in the art through experimental procedures.

[0102] By minimizing the operating cost through this objective function, emissions can reduce furnace-side consumption costs and improve the combustion rate of the fed fuel. At the same time, minimizing the predicted output power of the secondary reheat unit and the actual output power of the secondary reheat unit at time t can improve the control of insufficient power demand for the load in this embodiment, understand the amount of load demand in advance, and realize the rapid response of the secondary reheat unit to the grid and load demand.

[0103] S204. Integrate the constructed objective function and corresponding constraints into the multivariate predictive control model for training to obtain the trained multivariate predictive control model. Integrate the trained multivariate predictive control model into the furnace-side control subsystem. Collect furnace-side operating data in real time through sensors and input the real-time collected furnace-side operating data into the multivariate predictive control model to obtain the initial control variable parameters of the furnace side. Further, in this embodiment, the initial control variable parameters of the furnace side include fuel flow rate, wind speed, furnace steam pressure, excess air coefficient, and furnace steam temperature.

[0104] Furthermore, the specific steps for obtaining the initial peak-shaving limit in this embodiment include:

[0105] S401. Collect turbine-side attribute data and historical operating data and preprocess the collected data. Turbine-side attribute data includes the maximum threshold of output power of the secondary reheat unit, minimum steady-state load, start-up and shutdown losses, transient response capability, safety margin, steam parameters under design conditions (such as pressure and temperature), material properties of turbine blades, and design parameters of the flow passage (such as flow area and blade shape). Historical operating data includes rated power during actual operation, steam parameter fluctuation range, operating efficiency, start-up and shutdown records, and turbine performance test data under different loads.

[0106] S402. Construct a turbine peak-shaving interval prediction model using a neural network, and set an initial peak-shaving interval corresponding to the preliminary peak-shaving limit. The specific formula for the initial peak-shaving interval is as follows:

[0107] in, Indicates the initial peak-shaving interval, f a This represents the safety margin coefficient, a coefficient less than 1 set in actual operation to allow for contingencies in case of unforeseen events. In this embodiment, the value is between 0.8 and 0.9. This represents the minimum output power at which the steam turbine can operate stably at time t. This data was obtained by those skilled in the art through experimental procedures. This represents the additional energy loss caused by the turbine starting from a stopped state to full load or shutting down from full load at time t. This indicates the amount of output power that a steam turbine can increase or decrease per unit time.

[0108] S403. Input the preprocessed machine-side attribute data and historical operation data into the peak shaving interval prediction model for training, and obtain the trained peak interval prediction model.

[0109] S404. Input the real-time collected machine-side operating data and attribute data into the peak-shaving interval prediction model to predict the parameters corresponding to the peak-shaving interval, and use the predicted parameters to calculate the specific peak-shaving interval at the corresponding time; in this embodiment, the corresponding parameters include and

[0110] Furthermore, the specific steps for obtaining the initial control variable parameters on the machine side in this embodiment include:

[0111] S405. Collect and preprocess machine-side operating data and attribute data, and set the steam valve opening, cooling water volume, and extraction steam pressure setpoints as machine-side control variables;

[0112] S405. Based on all possible combinations of control variables from the current operating state to the initial peak-shaving limit, set the search space of the parameter search model constructed based on the particle swarm optimization algorithm, and simultaneously set the turbine reaction objective function and constraints, specifically as follows: J r This represents the objective function for improving the machine-side operating efficiency. This represents the difference between the current output power of the secondary reheat unit and the previous output power of the secondary reheat unit. Indicates the corresponding time of change, e f This represents the efficiency of the steam turbine. This value is obtained by using historical steam turbine operating efficiency data and a steam turbine efficiency prediction model built with support vector machines to predict the steam turbine efficiency data at the corresponding future time. w4 and w5 represent the corresponding weighting coefficients.

[0113] S406. Set the iteration cycle of the particle swarm optimization algorithm. Input the preprocessed machine-side running data and attribute data into the particle swarm optimization algorithm with the objective function and constraints set for training and search, and obtain the trained particle swarm parameter search model; in this embodiment, the iteration cycle is 200.

[0114] S407. The obtained particle swarm parameter search model is transferred from the machine side to the machine side control subsystem. The machine side operation data and attribute data collected in real time by the sensors are input into the particle swarm parameter search model to obtain the machine side initial control variable parameters.

[0115] Example 2

[0116] Please refer to Figure 4. Another embodiment of the present invention: a control system for rapid response to grid demand of a double reheat unit, including: a data processing module, a boiler-side control module, a turbine-side control module, a wind and solar power output prediction module, and a feedback control module.

[0117] The data processing module is used to collect monitoring data and load demand power data from the furnace side and the machine side using integrated sensors, and to preprocess the collected data. The preprocessing process includes cleaning, format conversion and standardization.

[0118] The furnace-side control module is used to construct furnace-side control and compensation strategies based on furnace-side operating data to optimize furnace-side operating control parameters in order to improve efficiency and response speed. The furnace-side control module includes a multivariable control unit and a dynamic compensation unit.

[0119] The multivariable control unit is used to construct a multivariable predictive control model and obtain the initial control variable parameters of the furnace side based on the acquired furnace side operating data and control strategy. The dynamic compensation unit is used to construct a dynamic compensation strategy, calculate the real-time heat flow distribution information of the furnace side based on the acquired furnace internal temperature field information and compensation strategy, and use the heat flow distribution information to compensate and correct the initial control variable parameters of the furnace side.

[0120] The turbine-side control module is used to optimize the turbine control parameters based on the acquired turbine-side operating data to meet peak-shaving requirements. The turbine-side control module includes a peak-shaving assessment unit and a parameter calculation unit.

[0121] The peak shaving assessment unit is used to comprehensively assess the steam turbine based on the acquired turbine-side attribute data and an assessment algorithm to obtain the initial peak shaving limit of the steam turbine; the parameter calculation unit is used to obtain the initial control variable parameters of the turbine side based on the acquired initial peak shaving limit and turbine-side operating data through the particle swarm algorithm.

[0122] The wind and solar power output prediction module is used to build and train the wind and solar power output prediction model. Based on the trained model, the predicted wind and solar power output data is calculated. The output power of the secondary reheat unit is calculated by combining the predicted wind and solar power output data with the load demand power data.

[0123] The feedback control module is used to optimize the control strategies on the furnace side and the turbine side based on the obtained output power of the secondary reheat unit, and to control the secondary unit to output response quickly; the feedback control module includes a parameter feedback optimization unit and a discrimination control unit;

[0124] The parameter feedback optimization unit is used to optimize the initial control variable parameters on the furnace side and the turbine side based on the obtained output power of the secondary reheat unit, so as to obtain the pre-control variable parameters on the furnace side and the turbine side. The discrimination control unit is used to set the output threshold of the secondary reheat unit and compare the threshold with the obtained output power of the secondary reheat unit to determine whether to generate electricity. If to generate electricity, the unit is controlled to generate electricity using the obtained pre-control variable parameters on the furnace side and the turbine side, and outputs a power value equal to the output power of the secondary reheat unit to make up for the power loss of wind and solar power.

[0125] Example 3

[0126] A computer-readable storage medium storing computer instructions that, when executed, provide a control method for a secondary reheat unit to rapidly respond to grid demands.

[0127] An electronic device includes a memory and a processor, the memory storing a computer program, and a control method for a secondary reheat unit to quickly respond to grid demands when the processor executes the computer program.

[0128] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the claims. All of these variations are within the protection scope of the present invention.

Claims

1. A control method for rapid response to grid demand in a double reheat unit, characterized in that, The specific steps include: Step S1: Acquire boiler-side and turbine-side attribute data, monitoring operation data and load demand power data of the secondary reheat unit through integrated sensors, and preprocess the collected data; Step S2: Construct a multivariate predictive control model. Based on the preprocessed furnace-side operating data and the multivariate predictive control model, obtain the initial control variable parameters of the furnace side. Step S3: Introduce a dynamic compensation strategy to obtain the internal temperature field of the furnace. Based on the obtained internal temperature field of the furnace and the dynamic compensation strategy, obtain real-time heat flow distribution information on the furnace side. Feed the obtained real-time heat flow distribution information back to the multivariate predictive control model to compensate and correct the obtained initial control variable parameters. Step S4: Introduce a turbine-side peak shaving strategy. Based on the turbine-side attribute data, use an evaluation algorithm to evaluate and obtain the initial peak shaving limit of the steam turbine. At the same time, based on the obtained initial peak shaving limit and turbine-side operating data, use a parameter search model constructed by the particle swarm optimization algorithm to obtain the initial control variable parameters of the turbine side. Step S5: Construct and train the wind and solar power output prediction model. Based on the acquired wind and solar power output data and load demand power data, use the wind and solar power output prediction model to calculate the output power of the secondary reheat unit. Step S6: Set the output threshold of the secondary reheat unit to 0, and feed back the obtained output power of the secondary reheat unit to the multivariate predictive control model and the turbine-side peak shaving strategy to obtain the furnace-side pre-control variable parameters and the turbine-side pre-control variable parameters. Step S7: Compare the output power of the secondary reheat unit with the output threshold of the secondary reheat unit. If it is greater than the threshold, control the secondary reheat unit using the obtained optimal control variable parameters of the pre-heater side and the optimal control variable parameters of the pre-heater side. Generate electricity according to the calculated output power of the secondary reheat unit and output a power value equal to the output power of the secondary reheat unit to make up for the power missing from wind and solar power output. If it is less than or equal to the threshold, the secondary reheat unit remains in a stopped state.

2. The control method for rapid response to grid demand of a double reheat unit as described in claim 1, characterized in that, The specific steps for constructing the wind and solar power output prediction model in step S5 include: S501. Collect historical wind and solar power output data, load balance power demand data, and weather factor data according to sunny, cloudy, overcast, and rainy weather conditions. Preprocess the collected data to construct a multi-weather-condition input sequence X. i =((x1…x n ), (y1…y m ), (z1…z k )); where X i Let x represent the data for the i-th weather condition. n y represents the nth wind and solar power output data in the i-th weather condition data. m z represents the m-th load balance power demand data in the i-th weather condition data. k This represents the k-th weather factor data in the i-th weather state data.

3. The control method for rapid response to grid demand of a double reheat unit as described in claim 2, characterized in that, The specific steps for constructing the wind and solar power output prediction model in step S5 also include: S502. Using the random forest algorithm, a wind and solar power output prediction model is constructed in an integrated manner, including a sunny prediction sub-model, a cloudy prediction sub-model, an overcast prediction model, and a rain prediction sub-model. S503, Input to multi-weather state input sequence X i =((x1…x n ), (y1…y m ), (z1…z k The weather conditions are input into the corresponding sub-models within the photovoltaic output prediction model for training. S504. Using historical wind and solar power output data and corresponding load balance power demand data, obtain historical true output error; at the same time, using wind and solar power output data and load balance power demand data predicted at corresponding historical times, obtain historical predicted output error. S504. Set the training threshold δ. Based on the historical actual output error and the historical predicted output error, obtain the secondary training error. Compare the obtained secondary training error with δ. If it is less than δ, the training is complete. Otherwise, continue training until the threshold condition is met.

4. The control method for rapid response to grid demand of a double reheat unit as described in claim 3, characterized in that, The specific calculation process for predicting the output power of the double reheat unit includes: S505. Collect wind and solar power output data, load balance power demand data and weather factor data for the first 30 hours of the forecast day, and input the wind and solar power output data and weather factor data for the first 30 hours into the trained wind and solar power output forecast model, and output wind and solar power output data for the next 12 hours. S506. Based on the wind and solar power output data and load balance power demand data for the next 12 hours, the output power of the secondary reheat unit is calculated using the load demand balance equation. The load demand balance equation is as follows: in, This represents the power required to balance the load at time t. This represents the predicted power output data of wind and solar power at time t. This represents the output power of the secondary reheat unit at time t.

5. The control method for rapid response to grid demand of a double reheat unit as described in claim 4, characterized in that, The specific steps of step S3 include: S301. Construct a three-dimensional geometric model of the furnace side using furnace side attribute data, divide the three-dimensional geometric model of the furnace side into N monitoring areas using a mesh algorithm, and obtain the temperature distribution data of the N monitoring areas inside the furnace through deployed furnace side sensors. S302. Based on the acquired temperature distribution data, the heat flow distribution map of each region in the furnace is calculated by combining finite element analysis with thermal radiation and hydrodynamic algorithms. S303. Feed back the obtained heat flow distribution map of each region to the initial control variable parameters, use the dynamic temperature compensation formula to calculate the correction value of the initial control variable parameters, and compensate and correct the initial control variable parameters.

6. The control method for rapid response to grid demand of a double reheat unit as described in claim 5, characterized in that, The dynamic temperature compensation formula is as follows: in, α represents the compensation amount for the i-th parameter of the initial control variable. t T represents the dynamic adjustment factor of the parameter. j T represents the real-time temperature data corresponding to the j-th monitoring area inside the furnace. rj K represents the expected temperature data corresponding to the j-th monitoring area inside the furnace, N represents the number of monitoring areas inside the furnace, and K represents the number of monitoring areas inside the furnace. p Indicates temperature deviation T j -T rj The corresponding correction parameter, K d Represents the rate of temperature change The corresponding correction parameter, K i Indicates cumulative temperature deviation The corresponding correction parameter, t0, represents the length of time it takes for the secondary reheat turbine unit to operate once.

7. The control method for rapid response to grid demand of a double reheat unit as described in claim 6, characterized in that, The specific steps for constructing the multivariate predictive control model include: S201. Preprocess the furnace-side attribute data, operating data, and load demand power data acquired by the integrated sensors to obtain furnace-side control variable data; S202. Construct a multivariate predictive control model using support vector machines, and input the acquired furnace-side control variable data into the multivariate predictive control model for training. S203. Set the objective function and corresponding constraints of the multivariate predictive control model to limit the furnace-side operating cost, operating efficiency and emissions, so as to minimize fuel consumption while maintaining the output power to meet the load power demand.

8. The control method for rapid response to grid demand of a double reheat unit as described in claim 7, characterized in that, The specific steps for constructing the multivariate predictive control model also include: S204. Integrate the constructed objective function and corresponding constraints into the multivariate predictive control model for training to obtain the trained multivariate predictive control model. S205. Integrate the trained multivariate predictive control model into the furnace-side control subsystem, collect furnace-side operating data in real time through sensors, and input the real-time collected furnace-side operating data into the multivariate predictive control model to obtain the initial control variable parameters of the furnace side.

9. The control method for rapid response to grid demand of a double reheat unit as described in claim 8, characterized in that, The objective function and corresponding constraints in S203 specifically include: The specific formula for the objective function is as follows: The constraints are: c min <c t <c max ,f min <f t <f max ,E0<E l , Where J represents the furnace-side objective constraint function. E0 represents the fuel consumption cost for one furnace operation cycle, and E0 represents the total pollution emissions corresponding to one furnace operation cycle. c represents the actual output power of the double reheat unit at time t; t c represents the fuel consumption at time t. min c max f represents the maximum and minimum fuel consumption per unit time, respectively. t f represents the air velocity fed into the furnace at time t. min f max These represent the maximum and minimum values ​​of the air velocity fed into the furnace at time t, respectively. E represents the maximum power output threshold of the double reheat unit at time t. l The threshold represents the total amount of pollution emissions during one cycle of furnace operation, and w1, w2, and w3 represent the corresponding weighting coefficients.

10. The control method for rapid response to grid demand of a double reheat unit as described in claim 9, characterized in that, The specific formula for the initial peak-shaving interval corresponding to the preliminary peak-shaving limit in step S4 is as follows: in, Indicates the initial peak-shaving interval, f a This represents the safety margin coefficient. This represents the minimum output power at which the steam turbine can operate stably at time t. This represents the additional energy loss caused by the turbine starting from a stopped state to full load or shutting down from full load at time t. This indicates the amount of output power that a steam turbine can increase or decrease per unit time.

11. The control method for rapid response to grid demand of a double reheat unit as described in claim 10, characterized in that, In step S4, the parameter search model sets the turbine reaction objective function and constraints, specifically: J r This represents the objective function for improving the machine-side operating efficiency. This represents the difference between the current output power of the secondary reheat unit and the previous output power of the secondary reheat unit. Indicates the corresponding time of change, e f This represents the efficiency of the steam turbine operation. This value is predicted using a steam turbine efficiency prediction model constructed with support vector machines based on the historical operating data of the steam turbine; w4 and w5 represent the corresponding weighting coefficients.

12. A control system for rapid response to grid demand of a double reheat unit, implemented based on the control method for rapid response to grid demand of a double reheat unit as described in any one of claims 1-11, characterized in that, include: Data processing module, furnace-side control module, turbine-side control module, wind and solar power output prediction module, and feedback control module; The data processing module is used to collect monitoring data and load demand power data from the furnace side and the machine side using integrated sensors, and to preprocess the collected data. The wind and solar power output prediction module is used to construct and train a wind and solar power output prediction model, calculate and predict wind and solar power output data based on the trained model, and calculate the output power of the secondary reheat unit by using the predicted wind and solar power output data and load demand power data.

13. The control system for rapid response to grid demand of a double reheat unit as described in claim 12, characterized in that, The furnace-side control module includes a multivariable control unit and a dynamic compensation unit; The multivariable control unit is used to construct a multivariable predictive control model and obtain initial control variable parameters of the furnace side based on the acquired furnace side operating data and control strategy. The dynamic compensation unit is used to construct a dynamic compensation strategy, calculate the real-time heat flow distribution information of the furnace side based on the acquired furnace internal temperature field information and compensation strategy, and use the heat flow distribution information to compensate and correct the initial control variable parameters of the furnace side.

14. The control system for rapid response to grid demand of a double reheat unit as described in claim 13, characterized in that, The machine-side control module includes a peak shaving assessment unit and a parameter calculation unit; The peak-shaving assessment unit is used to comprehensively assess the steam turbine based on the acquired turbine-side attribute data and an assessment algorithm to obtain the initial peak-shaving limit of the steam turbine; the parameter calculation unit is used to obtain the initial control variable parameters of the turbine side based on the acquired initial peak-shaving limit and the turbine-side operating data through a particle swarm optimization algorithm.

15. The control system for rapid response to grid demand of a double reheat unit as described in claim 14, characterized in that, The feedback control module includes a parameter feedback optimization unit and a discrimination control unit; The parameter feedback optimization unit is used to perform feedback optimization on the initial control variable parameters on the furnace side and the initial control variable parameters on the turbine side based on the obtained output power of the secondary reheat unit, so as to obtain the pre-control variable parameters on the furnace side and the pre-control variable parameters on the turbine side. The discrimination control unit is used to set the output threshold of the secondary reheat unit, and compare the threshold with the obtained output power of the secondary reheat unit to determine whether to generate electricity. If to generate electricity, the unit is controlled to generate electricity using the obtained pre-control variable parameters on the furnace side and the pre-control variable parameters on the turbine side, and outputs a power value equal to the output power of the secondary reheat unit.

16. A computer-readable storage medium, characterized in that, It stores computer instructions, which, when executed, perform the control method for rapid response to grid demand of the secondary reheat unit as described in any one of claims 1-11.

17. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the control method for rapid response to grid demand of a double reheat unit as described in any one of claims 1-11.

Images