Thermal power generating unit automatic cruise control method and system based on main steam temperature multi-stage linkage

Abstract

The embodiment of the invention provides a thermal power generating unit automatic cruise control method and system based on main steam temperature multi-stage linkage, and the method comprises the steps: abstracting a main steam temperature system as a controlled object model, and employing a first-order inertia and time delay transfer function as a mathematical structure of the controlled object model; characteristic parameters of the controlled object model are determined through parameter identification, and a main steam temperature dynamic characteristic simulation result is output; processing the input by adopting variational mode decomposition to obtain a plurality of intrinsic mode components; extracting a ground state component representing a long-term change trend and a dynamic component representing a short-term change trend from the intrinsic mode component; executing forward cruise calculation and reverse cruise calculation, and outputting a calculation result of forward cruise and a calculation result of reverse cruise; and performing smoothing and amplitude limiting processing on the calculation result of the forward cruise and the calculation result of the reverse cruise through an activation function, and outputting an automatic adjustment instruction.

Description

Technical Field

[0001] This application relates to the technical field of industrial automatic control, specifically to an automatic cruise control method and system for thermal power units based on multi-level linkage of main steam temperature. Background Technology

[0002] The desuperheating valve is a crucial component of the main steam temperature control system in thermal power units. It atomizes desuperheating water and injects it directly into the steam stream. The desuperheating effect is controlled by adjusting the valve opening to regulate the amount of injected desuperheating water. The effectiveness of the desuperheating valve significantly impacts the quality of main steam temperature regulation. If the desuperheating valve operates too quickly, it will cause fluctuations in the main steam temperature; if it operates too slowly, it will expose the main steam to the risk of overheating.

[0003] With the construction of new power systems, thermal power plants require frequent peak and frequency regulation, leading to significant fluctuations in main steam temperature. This places increasingly higher demands on the accuracy and speed of desuperheating valve position control. However, current research and applications of desuperheating valve control systems, both domestically and internationally, are based on the unit's normal operating conditions, where operating parameters are stable, nonlinearity and time delay are minimal, and disturbances are relatively few. However, under deep peak and frequency regulation conditions, the spatial and temporal distribution of heat is uneven, disturbances are more multi-source and complex, and main steam temperature fluctuations are more severe.

[0004] Existing technologies utilize the excellent disturbance rejection capabilities and nonlinear adaptability of active disturbance rejection algorithms to improve the control quality of main steam temperature. However, active disturbance rejection controllers are not highly adaptable to objects with large time delays, requiring predictive control algorithms to compensate, making the system dependent on a relatively accurate model. Furthermore, this method aims to control the main steam temperature at a given setpoint, but does not discuss how to determine the setpoint; this method is only applicable to secondary desuperheaters.

[0005] In summary, current research on control systems for main steam temperature faces the following technical challenges: existing technologies focus on controlling single-stage desuperheating water, with insufficient development on how to coordinate and match the effects of two-stage desuperheating water; existing technologies are limited to optimizing control algorithms, with insufficient attention paid to setting the target value; and the adjustment of the set value relies on the experience of operators, which greatly increases the workload and labor intensity of operators during frequent peak and frequency adjustments. Summary of the Invention

[0006] This application proposes an automatic cruise control method and system for thermal power units based on multi-level linkage of main steam temperature, in order to overcome the deficiencies of the prior art.

[0007] According to a first aspect of the embodiments of this application, an automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature is provided, comprising: The main steam temperature system is abstracted into a controlled object model, and a first-order inertial plus time delay transfer function is used as the mathematical structure of the controlled object model. The characteristic parameters of the controlled object model are determined by parameter identification, and the simulation results of the main steam temperature dynamic characteristics are output. The simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal are used as inputs, and the inputs are processed by variational mode decomposition to obtain multiple intrinsic mode components. The ground state component, which represents the long-term trend, and the dynamic component, which represents the short-term trend, are extracted from the intrinsic modal components. Using the ground state component and the dynamic component as core inputs, forward cruise calculation and reverse cruise calculation are performed, and the calculation results of forward cruise and reverse cruise are output. The calculation results of the forward cruise and the reverse cruise are smoothed and limited by the activation function, and an automatic adjustment command is output.

[0008] In some embodiments, the method further includes: Water-coal ratio adjustment is used as a coarse adjustment method to respond to adjustment commands for intermediate point temperature setpoints; The primary desuperheating water regulation is used as a fine-tuning method to respond to the adjustment command of the secondary superheater outlet steam temperature setpoint; The system uses secondary desuperheating water regulation as a fine-tuning method to respond to adjustment commands for the main steam temperature setpoint.

[0009] In some implementations, the transfer function model includes three undetermined parameters: gain, time constant, and lag time. The determination of the characteristic parameters of the controlled object model through parameter identification includes: A step disturbance signal is applied to the desuperheating valve, and the real-time response data of the main steam temperature is recorded; The particle swarm optimization algorithm is used to iteratively optimize the gain, the time constant, and the lag time parameter until the fit between the output of the controlled object model and the real-time response data of the main steam temperature meets a preset threshold.

[0010] In some implementations, the step of using the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal as input, and processing the input using variational mode decomposition to obtain multiple intrinsic mode components includes: The preset number of modal decompositions and the range of values ​​for the penalty parameter are defined. Using the minimum envelope entropy as the fitness function, an optimization algorithm is used to determine the target combination of the number of mode decompositions K and the penalty parameters. The input signal is decomposed using the target combination to obtain multiple intrinsic mode components; The number of modal decompositions ranges from 3 to 8, and the penalty parameter ranges from 100 to 3000.

[0011] In some implementations, extracting the ground-state component representing the long-term trend and the dynamic component representing the short-term trend from the intrinsic modal components includes: The intrinsic mode component with the lowest frequency is identified as the ground state component; The intrinsic modal components whose frequency matches the operating cycle of the desuperheating valve are identified as the dynamic components; The intrinsic mode components at other frequencies are identified as noise signals and discarded.

[0012] In some implementations, performing forward cruise calculations includes: Based on the dynamic component of the intermediate point temperature, the first adjustment amount of the secondary superheater outlet steam temperature setpoint is calculated and obtained; The first adjustment amount is combined with the dynamic component of the secondary superheater outlet steam temperature to calculate the second adjustment amount of the main steam temperature setpoint.

[0013] In some implementations, performing the reverse cruise calculation includes: Calculate the deviation between the ground state component and the main steam temperature setpoint, and calculate the deviation between the actual opening degree of the desuperheating valve and the target opening degree; The deviation between the ground state component and the main steam temperature setpoint, as well as the deviation between the actual opening degree and the target opening degree of the desuperheating valve, are fused and calculated using a preset conversion function to obtain the third adjustment amount of the secondary superheater outlet steam temperature setpoint. Based on the third adjustment amount and the deviation between the actual opening degree and the target opening degree of the desuperheating valve, an intermediate adjustment amount for the intermediate point temperature setpoint is generated.

[0014] In some implementations, the activation function is a modified hyperbolic tangent function used to simulate the operational inertia of operators; wherein, the mathematical expression of the activation function includes: the output automatic adjustment command is zero when it is in the dead zone, the output automatic adjustment command is a constant value when it is in the saturation zone, and the output automatic adjustment command exhibits an S-shaped curve when it is in the linear zone.

[0015] In some implementations, the parameters of the activation function include the dead zone and the saturation region, wherein adjusting the dead zone range and the value of the saturation region includes: The dead zone range is adjusted by defining the upper limit of the positive cutoff region and the lower limit of the negative cutoff region; By defining a positive saturation upper limit and a negative saturation lower limit, the constant value of the saturation region is adjusted.

[0016] According to a second aspect of this application, an automatic cruise control system for thermal power units based on multi-level linkage of main steam temperature is provided, comprising: The model building module is used to abstract the main steam temperature system into a controlled object model, and adopts a first-order inertial plus time delay transfer function as the mathematical structure of the controlled object model. The simulation result output module is used to determine the characteristic parameters of the controlled object model through parameter identification and output the simulation results of the main steam temperature dynamic characteristics. The intrinsic mode generation module is used to take the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal as input, and process the input using variational mode decomposition to obtain multiple intrinsic mode components. The modal component extraction module is used to extract the ground state component, which represents the long-term trend, and the dynamic component, which represents the short-term trend, from the intrinsic modal components. The cruise calculation module is used to perform forward cruise calculation and reverse cruise calculation with the ground state component and the dynamic component as core inputs, and output the calculation results of forward cruise and reverse cruise. The adjustment instruction generation module is used to smooth and limit the calculation results of the forward cruise and the reverse cruise through an activation function, and output an automatic adjustment instruction.

[0017] The beneficial effects of the automatic cruise control method and system for thermal power units based on multi-level linkage of main steam temperature in the embodiments of this application include at least the following: This application's embodiments abstract the main steam temperature system into a controlled object model, simplifying the complex, nonlinear actual thermodynamic system into a linearized model with a well-defined structure and identifiable parameters. This provides a unified and computable object foundation for all subsequent quantitative analyses and algorithm designs, making model-based control strategy design possible and overcoming the limitations of relying entirely on field tests and expert experience. Through parameter identification technology, the general model structure can accurately match the dynamic characteristics of a specific unit, ensuring that the "dynamic characteristic simulation results" output by the model truly reflect the unit's response under disturbances. This provides high-quality, representative input signals for subsequent feature decomposition, improving the adaptability and accuracy of the entire method for specific objects. By utilizing variational mode decomposition, an advanced signal processing method, the mixed steam temperature signal can be adaptively and nearly orthogonally decomposed into eigenmode components of different frequency scales, achieving preliminary separation of components with different physical meanings in the original signal. This lays the foundation for accurately extracting control-related feature quantities and solves the problem of... Traditional methods struggle to effectively separate components with different trends in a signal. By extracting ground-state and dynamic components, this method physically distinguishes long-term trends caused by slow load changes from short-term fluctuations caused by the action of a desuperheating valve. This allows the controller to adopt different strategies for changes at different time scales, providing information for refined and proactive control. Through forward and reverse cruise calculations, proactive and anticipatory adjustments are achieved based on dynamic components, reserving space for temperature rise and effectively preventing overheating. Rebalancing of heat distribution is achieved based on ground-state components, correcting energy imbalances at their source. This enables automatic and coordinated optimization of multi-level setpoints, effectively solving the problems of untimely and uncoordinated manual adjustments. Activation functions ensure the stability and safety of control commands, avoiding frequent setpoint oscillations caused by minor signal fluctuations and improving system stability. The "saturation zone" characteristic limits the maximum amplitude of a single adjustment, preventing over-adjustment and ensuring a smooth process. This makes the behavior of the automatic control system more like that of a cautious operator, enhancing the system's robustness and engineering practicality. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature, according to an embodiment of this application. Figure 2 This is a flow chart of the steam and water flow of the steam temperature system according to an embodiment of this application. Figure 3 The main steam temperature and IMF1-3 obtained from VMD decomposition in the embodiments of this application are shown. Figure 4 This is a schematic diagram of the f2 function in an embodiment of this application; Figure 5 This is a schematic diagram of the cruise simulation results of the superheat setpoint in an embodiment of this application; Figure 6 This is a schematic diagram of the cruise simulation results of the secondary superheater outlet temperature setpoint in an embodiment of this application; Figure 7 This is a schematic diagram of the main steam temperature setpoint cruise simulation results in an embodiment of this application; Figure 8 This is a cruise chart of the setpoint of the outlet steam temperature of the secondary superheater of a 1000MW unit according to an embodiment of this application; Figure 9 This is a cruise chart of the main steam temperature setpoint for a 1000MW unit according to an embodiment of this application; Figure 10 This is a schematic diagram of the automatic cruise control system for thermal power units based on multi-level linkage of main steam temperature, according to an embodiment of this application. Figure 11 This is a schematic diagram of the linkage control system based on constant value cruise according to an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the automatic cruise control method and system for thermal power units based on multi-level linkage of main steam temperature will be described clearly and completely below in conjunction with the accompanying drawings of the embodiments of this application. Obviously, the described embodiments are only some embodiments of the embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed embodiments of the present application, but merely to illustrate selected embodiments of the present application. Other embodiments obtained by those skilled in the art based on the embodiments of the present application without inventive effort are all within the scope of protection of the embodiments of the present application.

[0021] It can be noted that similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it will not be further defined and explained in subsequent figures according to the embodiments of this application.

[0022] This application discloses an automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature. The method is executed based on an automatic cruise control system for thermal power units based on multi-level linkage of main steam temperature, and includes steps 110-160.

[0023] Step 110: The main steam temperature system is abstracted into a controlled object model, and a first-order inertial plus time delay transfer function is used as the mathematical structure of the controlled object model.

[0024] In some implementations, the transfer function model includes three undetermined parameters: gain, time constant, and lag time.

[0025] In some implementations, the method further includes: based on the water-to-coal ratio ( f w ） The adjustment is used as a coarse adjustment method to respond to the adjustment command of the intermediate point temperature setpoint; the adjustment based on the first-stage desuperheating water is used as a fine adjustment method to respond to the adjustment command of the second-stage superheater outlet steam temperature setpoint; and the adjustment based on the second-stage desuperheating water is used as a fine adjustment method to respond to the adjustment command of the main steam temperature setpoint.

[0026] For example, refer to the appendix. Figure 2 The diagram shown illustrates the steam-water flow path of the steam temperature system in an ultra-supercritical unit. Figure 2 As shown, This refers to the outlet temperature of the first-stage superheater. This refers to the outlet temperature of the first-stage desuperheater. This refers to the outlet temperature of the secondary superheater. This refers to the outlet temperature of the secondary desuperheater. Main steam temperature, This refers to the valve opening of the first-stage desuperheater. For the valve opening of the secondary desuperheater, The main steam flow rate disturbance. For the disturbance of smoke, The main steam temperature is primarily affected by steam flow rate, flue gas heat, and desuperheating water flow rate. Steam flow rate and flue gas heat are the main external disturbances, while desuperheating water flow rate is the main controlled variable. The main steam temperature control employs a multi-stage control method, including intermediate point temperature and first-stage desuperheater valve opening. and the opening degree of the secondary desuperheater valve Joint regulation, for example, by the water-to-coal ratio ( ） Perform coarse adjustment, adjusting the opening of the primary desuperheater valve. Fine-tuning was performed on the opening of the secondary desuperheater valve. Fine-tuning is performed. The water-coal ratio, primary desuperheating water, and secondary desuperheating water are adjusted to correspond to coarse, fine, and precise adjustments, respectively, achieving three-level linkage.

[0027] In some relevant embodiments of this application, typically, given and The setting value is and The difference between the two (i.e., the deviation) , This is the controlled variable of the desuperheater. For example, , The calculation method is shown in the following equations (1) and (2): Equation (1); Equation (2); Among them, respectively adopt and To overcome the two deviations mentioned above , This allows for decoupled control, with each desuperheater's control object clearly defined, facilitating understanding and intervention by operators. However, in new power systems, units require frequent deep peak shaving and rapid frequency regulation, leading to decreased boiler combustion stability, insufficient hydrodynamics, and migration of radiative / convective heat transfer characteristics, potentially resulting in uneven spatial and temporal heat distribution, such as... and Reverse, causing and Reverse action. Due to Design traffic is high. The design traffic is small, and in extreme cases, it may occur. The opening is small, and When fully open, the settings need to be adjusted promptly. However, and Adjustments rely heavily on the experience of operators, and during frequent peak and frequency adjustments, the workload and labor intensity of operators are greatly increased.

[0028] Step 120: Determine the characteristic parameters of the controlled object model through parameter identification, and output the simulation results of the main steam temperature dynamic characteristics.

[0029] The simulation results of the main steam temperature dynamic characteristics are used as the input for the variational mode decomposition in step 130.

[0030] In some implementations, the determination of the characteristic parameters of the controlled object model through parameter identification includes: applying a step disturbance signal to the desuperheating valve and recording the real-time response data of the main steam temperature; using a particle swarm optimization algorithm to iteratively optimize the gain, the time constant, and the lag time parameter until the fit between the output of the controlled object model and the real-time response data of the main steam temperature meets a preset threshold.

[0031] Among them, the particle swarm optimization algorithm is used to minimize the error between the model output and the real-time response data, ensuring parameter accuracy.

[0032] For example, according to the appendix Figure 2 In the embodiment, without considering disturbances, the extraction from to The transfer function of the soft drink process is given by equation (3) below: Equation (3); in, It is the complex frequency variable in the Laplace transform, and is usually called the Laplace operator or complex variable.

[0033] Furthermore, without considering disturbances, the extraction from to The transfer function of the soft drink process is given by equation (4) below: Equation (4); Among these, the steam flow rate disturbance may be transmitted step by step, first affecting... However, due to the counter-current flow of flue gas, flow disturbances can be mitigated without affecting... Direct impact in the case Then we get equations (5) and (6) respectively, as shown below: Equation (5); Equation (6); From equations (5) and (6), the overall steam-water flow transfer function is shown below: Equation (7); For example, by combining equations (1) and (5) above, and equations (2) and (7) above, we can obtain equations (8) and (9) as follows: Equation (8); Equation (9); This is the theoretical calculation value of the deviation in equations (1) and (2), which includes multiple signals. These multiple signals refer to the signals that constitute the two deviations. , The theoretical calculation model is a set of all signals that fully describe the inputs, outputs, targets, disturbances, and internal states of the control system.

[0034] Step 130: Using the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal as input, the input is processed by variational mode decomposition to obtain multiple intrinsic mode components (hereinafter referred to as "IMF components").

[0035] In some implementations, the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal are used as inputs, and the inputs are processed by variational mode decomposition to obtain multiple intrinsic mode components, including: a preset number of mode decompositions. The range of values ​​for the penalty parameter α; using the minimum envelope entropy as the fitness function, the number of mode decompositions is determined through an optimization algorithm. The target combination of the penalty parameter is used to decompose the input signal to obtain multiple intrinsic mode components; wherein the number of mode decompositions K ranges from 3 to 8, and the penalty parameter α ranges from 100 to 3000.

[0036] in, The values ​​of 3-8 and α, ranging from 100-3000, were determined based on preliminary experiments to balance decomposition accuracy and computational efficiency.

[0037] For example, the input can be processed by variational mode decomposition (VMD) based on the feature decomposition method, which can further optimize the structure and mode of this application by fully extracting the effective signal features of formulas (8) and (9).

[0038] Among related technologies, Empirical Mode Decomposition (EMD) is suitable for the analysis and processing of nonlinear and non-stationary signals. It has high requirements for the non-stationarity of data, but it is sensitive to noise and suffers from mode aliasing. Ensemble Empirical Mode Decomposition (EEMD) achieves auxiliary decomposition by adding Gaussian white noise to the original signal, which can effectively make up for the inherent mode aliasing problem in the non-smooth time series decomposition of EMD. However, it greatly increases the computational complexity.

[0039] Variational mode decomposition is a nonlinear, adaptive signal decomposition method with good feature extraction capabilities.

[0040] For example, this application constructs a variational mode decomposition parameter optimization model based on minimum average envelope entropy and optimizes it using the marine predator algorithm. Comparison with no decomposition and empirical mode decomposition demonstrates the model's optimal performance. Variational mode decomposition can decompose non-stationary coal-fired power plant data sequences into a finite number of intrinsic mode components and residual components, better handling complex nonlinear and non-stationary signals and improving the accuracy and effectiveness of coal-fired power plant modeling. The calculation formula for the variational mode decomposition algorithm of this application is shown below: Equation (10); Equation (11); in, The result is obtained based on variational mode decomposition. One IMF component, The center frequency of each IMF component. The imaginary unit is used to construct the Hilbert transform kernel function in variational mode decomposition algorithms; This is a time variable used to indicate that the algorithm model is defined in the continuous time domain; The original input signal specifically refers to the non-stationary coal-fired power plant data sequence (such as steam temperature signal) that needs to be decomposed.

[0041] Among them, the number of modes and penalty parameters These are adjustable parameters, and both determine the signal quality. Number of modes The number of IMF components affects the decomposition process; too many components lead to over-decomposition, while too few lead to under-decomposition. (Penalty parameter) The bandwidth that affects IMF components is too high, leading to information loss, while too low bandwidth results in information redundancy.

[24] Based on the variational mode decomposition algorithm, this application uses the minimum envelope entropy as the fitness function of the Golden Jackal optimization algorithm to determine the optimal parameters and optimal eigenvalues ​​of the superheated steam temperature object based on the variational mode decomposition algorithm.

[0042] Table 1: Optimal VMD parameters and eigenvalues ​​for steam temperature objects

[0043] Referring to the above description, in a specific embodiment of this application, 1800 data points were generated over 30 minutes during a deep adjustment of a 1000MW ultra-supercritical coal-fired power unit. and The data were decomposed using the variational mode decomposition algorithm with the parameters in Table 1. At the same time, the root mean square error (RMSE) of the reproduction results was statistically analyzed, and the results are shown in Table 3 below.

[0044] Table 3: RMSE Statistics of VMD Decomposition Results for a Certain Ultra-Supercritical Unit

[0045] The reproduction results show that, using the parameters in Table 2, the variational mode decomposition algorithm decomposes the data. IMF1-7 signal summation and The summation of the IMF1-6 signals all reached a thousandths level similarity to the original signal. Among them, after the signal is decomposed to the IMF3 level, further decomposition has little improvement on RMSE, so IMF1-3 is sufficient.

[0046] Step 140: Extract the ground state component representing the long-term trend and the dynamic component representing the short-term trend from the intrinsic mode components.

[0047] The identification of ground state components and dynamic components is based on frequency filtering, such as IMF1 being the ground state and IMF2 being the dynamic component.

[0048] In some implementations, extracting the ground state component representing the long-term trend and the dynamic component representing the short-term trend from the intrinsic modal components includes: identifying the intrinsic modal component with the lowest frequency as the ground state component; identifying the intrinsic modal component whose frequency matches the operating cycle of the desuperheating valve as the dynamic component; and identifying the intrinsic modal components of other frequencies as noise signals and discarding them.

[0049] For example, with For example, refer to the appendix. Figure 3 As shown, IMF1-3 and The relationship. (By appendix) Figure 3 It can be seen that, The ground state component reflects The long-term trend of change; As a dynamic component, it reflects Short-term trends; The subsequent IMF signals are noise signals that can be ignored.

[0050] For the desuperheating valve Performing the same operation, we can obtain equations (12) and (13) as follows: Equation (12); Equation (13); The above calculation shows the correspondence between the ground state and dynamic components of the steam temperature and the corresponding ground state and dynamic components of the desuperheating valve opening.

[0051] Among them, the appendix Figure 3 of As The long-term trend of change is calculated by equation (6) shown above, which is ground state components and The ground-state component overcomes the long-term variation in flue gas flow rate caused by load changes. The sum of the effects of the low-frequency components yields the following equation (14): Equation (14); The period is approximately 5 minutes, and the amplitude is within ±10℃, which is similar to the period and range of the desuperheating water. It can be approximated as the steam temperature change caused by the periodic action of the desuperheating water. Equation (15) derived from equation (6) is shown below: Equation (15); The subsequent signal frequency is very high, while the amplitude is ≤0.1℃, which can be considered as the vibration of the measured signal. Noise signals generated by high-frequency components of flue gas disturbances, etc.

[0052] Similarly, for The decomposition results, combined with equation (5), yield the following equations (16) and (17): Equation (16); Equation (17).

[0053] Step 150: Using the ground state component and the dynamic component as core inputs, perform forward cruise calculation and reverse cruise calculation, and output the calculation results of forward cruise and reverse cruise.

[0054] Among them, setpoint cruise involves adjusting setpoints within a reasonable range based on unit operating conditions, achieving system self-adaptation in multiple modes. This can replace or reduce manual intervention, lowering the workload of operators. In this application, the object of setpoint cruise is... and The setpoint and intermediate point temperature setpoint are only suggested adjustments due to their involvement in water-to-coal ratio regulation. Applications are submitted in accordance with... As a representation of the intermediate point temperature, the setpoint cruise is more durable and causes less disturbance to the system compared to using feedforward regulation. The setpoint cruise in this application includes two directions: forward cruise and reverse cruise.

[0055] In some implementations, the forward cruise calculation includes: calculating and obtaining a first adjustment amount ΔT23 for the secondary superheater outlet steam temperature setpoint based on the dynamic component of the intermediate point temperature; and combining the first adjustment amount ΔT23 with the dynamic component of the secondary superheater outlet steam temperature to calculate a second adjustment amount ΔT42 for the main steam temperature setpoint.

[0056] In some embodiments, the reverse cruise calculation includes: calculating the deviation between the ground state component and the main steam temperature setpoint, and calculating the deviation between the actual opening degree of the desuperheating valve and the target opening degree; fusing and calculating the deviation between the ground state component and the main steam temperature setpoint and the deviation between the actual opening degree of the desuperheating valve and the target opening degree through a preset conversion function, and obtaining a third adjustment amount ΔT21 for the secondary superheater outlet steam temperature setpoint; and generating an intermediate adjustment amount ΔT01 for the intermediate point temperature setpoint based on the third adjustment amount ΔT21 and the deviation between the actual opening degree of the desuperheating valve and the target opening degree.

[0057] For example, reverse setpoint cruise refers to utilizing Temperature to adjust Set the value, then adjust the midpoint temperature setpoint, that is... → → The purpose is to adjust the water-coal balance and heat distribution by utilizing the ground-state components.

[0058] For example, ultra-supercritical units If the temperature is controlled at 600℃ (adjustable according to boiler design), then... The ground state component deviation is shown as follows: Equation (18); Desuperheater Excessive opening will reduce the regulation margin and decrease the unit's economic efficiency. This application optimizes... The ground-state component deviation is expressed as follows: Equation (19); in, When designing a boiler Ideal opening, The conversion function is shown in equation (20) below: Equation (20); in, To adjust the dead zone, This is the saturated conversion opening.

[0059] Based on the above equation (14), we can obtain The low-frequency part and The excess portion was factored in. The calculated ground-state component deviation is shown below: Equation (21); in, That is, when the cruise is in reverse setting. The adjustment amount.

[0060] Ultra-supercritical units The temperature needs to be controlled at 550℃ (adjustable according to boiler design). The ground state component deviation is shown as follows: Equation (22); The ground state component deviation is shown as follows: Equation (23); in, When designing a boiler Ideal opening, The conversion function is shown in equation (20).

[0061] Based on equation (17), we can obtain The low-frequency part and The excess portion was factored in. The calculated ground-state component deviation is shown as follows: Equation (24); in, Δ t 01 That is, when the cruise is in reverse setting. The adjustment amount.

[0062] For example, positive setpoint cruise refers to using the midpoint temperature to adjust... t 2. Set the value, then... t 4. Setting value, i.e. → → The purpose is to improve the quality of steam temperature control by anticipating the dynamic components.

[0063] According to equation (3), Dynamic components are transmitted to The location is shown as: Equation (25); in, That is, during positive setpoint cruise. The adjustment amount. The first adjustment amount in this embodiment corresponds to the above formula (25). The third adjustment amount is calculated from the dynamic components of the intermediate point temperature, and corresponds to the above equation (21). The intermediate adjustment amount corresponds to equation (24). It is obtained by fusing the bias through the conversion function.

[0064] According to equation (4), Dynamic components are transmitted to The location is shown as: Equation (26); in, That is, during positive setpoint cruise. The adjustment amount.

[0065] Step 160: Using an activation function, smooth and limit the calculation results of the forward cruise and the reverse cruise, and output an automatic adjustment command.

[0066] In some implementations, the activation function is an improved hyperbolic tangent function used to simulate the operational inertia of operators; wherein the mathematical expression of the activation function includes: the output automatic adjustment command is zero in the dead zone, the output automatic adjustment command is a constant value in the saturation zone, and the output automatic adjustment command exhibits an S-shaped curve in the linear zone.

[0067] In some implementations, the parameters of the activation function include the dead zone and the saturation zone. The adjustment of the dead zone range and the saturation zone value includes: defining an upper limit k2 for the positive cutoff zone and a lower limit k3 for the negative cutoff zone to prevent small fluctuations and adjust the dead zone range; and defining an upper limit k4 for the positive saturation zone and a lower limit k5 for the negative saturation zone to adjust the constant value of the saturation zone and limit the adjustment range.

[0068] For example, the forward and reverse cruise results are calculated as follows: Combining equations (21) and (25), we can obtain The adjustment amount is shown as follows: Equation (27); According to equation (26), we can obtain The adjustment amount is shown as follows: Equation (28); According to equation (24), we can obtain The adjustment amount is shown as follows: Equation (29); in, The activation function is used to prevent frequent adjustments of the set value and to mimic the operating characteristics of operators. In neural networks, common activation functions are shown in equations (30) and (31).

[0069] Sigmoid function: Equation (30); Tanh function: Equation (31); The modified equation (31) is shown as follows: Equation (32); in, 、 This represents the upper limit of the positive and negative cutoff zone. The upper limit of negative saturation. This represents the upper limit of positive saturation.

[0070] See attached document Figure 4 As shown, the activation function is illustrated. Coordinate graph, where, 、 , , Designed in accordance with the actual operating conditions of the unit, for example, = , k 4= k 5. 、 Used to influence cruise timing , Used to affect cruise amplitude.

[0071] This application's embodiments abstract the main steam temperature system into a controlled object model, simplifying the complex, nonlinear actual thermodynamic system into a linearized model with a well-defined structure and identifiable parameters. This provides a unified and computable object foundation for all subsequent quantitative analyses and algorithm designs, making model-based control strategy design possible and overcoming the limitations of relying entirely on field tests and expert experience. Through parameter identification technology, the general model structure can accurately match the dynamic characteristics of a specific unit, ensuring that the "dynamic characteristic simulation results" output by the model truly reflect the unit's response under disturbances. This provides high-quality, representative input signals for subsequent feature decomposition, improving the adaptability and accuracy of the entire method for specific objects. By utilizing variational mode decomposition, an advanced signal processing method, the mixed steam temperature signal can be adaptively and nearly orthogonally decomposed into eigenmode components of different frequency scales, achieving preliminary separation of components with different physical meanings in the original signal. This lays the foundation for accurately extracting control-related feature quantities and solves the problem of... Traditional methods struggle to effectively separate components with different trends in a signal. By extracting ground-state and dynamic components, this method physically distinguishes long-term trends caused by slow load changes from short-term fluctuations caused by the action of a desuperheating valve. This allows the controller to adopt different strategies for changes at different time scales, providing information for refined and proactive control. Through forward and reverse cruise calculations, proactive and anticipatory adjustments are achieved based on dynamic components, reserving space for temperature rise and effectively preventing overheating. Rebalancing of heat distribution is achieved based on ground-state components, correcting energy imbalances at their source. This enables automatic and coordinated optimization of multi-level setpoints, effectively solving the problems of untimely and uncoordinated manual adjustments. Activation functions ensure the stability and safety of control commands, avoiding frequent setpoint oscillations caused by minor signal fluctuations and improving system stability. The "saturation zone" characteristic limits the maximum amplitude of a single adjustment, preventing over-adjustment and ensuring a smooth process. This makes the behavior of the automatic control system more like that of a cautious operator, enhancing the system's robustness and engineering practicality.

[0072] The following is a detailed implementation process of an automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature.

[0073] Based on 1800 data points (recording period: seconds) from a 1000MW ultra-supercritical coal-fired power unit during peak shaving, simulation experiments were conducted at 750MW to test the cruise of the intermediate point temperature setpoint, the cruise of the secondary superheater outlet steam temperature setpoint, and the cruise of the main steam temperature setpoint. The experimental parameters are shown in Table 3 below. Table 3: Simulation k Value setting table

[0074] in, , and of k 2. k 3. k 4 and k 5 gradually decreases, corresponding to Perform a coarse adjustment. u 1. Make fine adjustments. The feature allows for fine-tuning. It is worth noting that... k The setpoint design needs to be finely adjusted based on the actual operating conditions of the unit. This article only provides a rough design to determine whether the timing and direction of the setpoint cruise are correct.

[0075] See attached document Figure 5 As shown in the figure, the superheat, its setpoint, and the load are the original data, while the cruise value is obtained from the simulation. The superheat is the difference between the midpoint temperature and the saturated steam temperature, and it can also characterize the water-coal balance and heat distribution. During the first half of the load increase process, the superheat did not deviate significantly from the setpoint, and the operators did not intervene. However, in the short period after switching to a load decrease, due to the decrease in load rate, the superheater's heat absorption decreased (the superheater is a convective heat exchanger, and its heat absorption capacity weakens as the load decreases). First to drop (compare with attached) Figure 5 Appendix Figure 6 and attached Figure 7 (It can be obtained), however, the superheat was not yet affected at this time, so the operators still did not intervene. The linkage algorithm, through reverse setpoint cruise, extracted this trend and adjusted the superheat setpoint in a timely manner (i.e., cruise 1), nearly 2 minutes earlier than the operator's manual operation 1. Furthermore, with the increased heat absorption of the secondary superheater (the secondary superheater is a radiative heat exchanger, whose heat absorption capacity increases as the load decreases) and u 1. The opening decreases. It began to rise rapidly (compare with the attached image) Figure 5 Appendix Figure 6 As can be seen, the linkage algorithm extracted this trend through reverse setpoint cruise and adjusted the superheat setpoint in a timely manner (i.e., cruise 2), reducing the setpoint before the superheat rapidly increased. The operators failed to reduce the superheat setpoint in time, resulting in excessively high superheat, which could easily lead to overheating of the water-cooled wall. Finally, although the superheat remained far above the setpoint, and The temperature was already far below the setpoint, and the linkage algorithm, through reverse setpoint cruise, extracted this trend and promptly adjusted the overheat setpoint (i.e., cruise 3), raising the setpoint to prevent the overheat from declining too quickly and reducing the unit's economy. Based on this, the above summary is shown in Table 4 below: Table 4: Summary of Superheat Setpoint Cruise

[0076] See attached document Figure 6 As shown, the outlet steam temperature of the second passage and setting value 1. Reduce opening degree The opening degree is the original data, and the cruise value is obtained from the simulation. Load changes and additional Figure 5 The same. During the first half of the load increase process, the operators manually reduced the opening by nearly 20 degrees using manual control 1. The linkage algorithm calculated at this time... The overheating trend has remained stable in the long term, and is insufficient to overcome. k 2. k 3 dead zones, no action taken. In the short period after switching to load reduction, due to the decrease in load factor, heat absorption decreases. It descended first, but at this time Initially affected, operators did not intervene. However, the linkage algorithm, through reverse setpoint cruise, extracted this trend and promptly adjusted the superheat setpoint (i.e., cruise 1), nearly 2 minutes earlier than the operator's manual operation 2. Furthermore, with the enhancement of secondary superheat absorption and... The valve is reduced; at this time, It began to rebound rapidly, and the linkage control algorithm quickly adjusted the setpoint (cruise 2). At this point, the operator's manual operation (3) clearly misjudged the situation. Finally, t and The water level dropped rapidly, and the linkage algorithm detected this trend and promptly adjusted the setpoint (i.e., cruise 3), increasing the setpoint to shut off the desuperheating water as quickly as possible. Based on this, the above summary is shown in Table 5 below: Table 5: Summary of Secondary Superheater Outlet Temperature Setpoint Cruise

[0077] See attached document Figure 7 As shown, main steam temperature and its settings The second reduction opening degree is u 2. Raw data, cruise values ​​are obtained from simulation, load changes and additional data. Figure 5 Same. During the first half of the load increase process, the operator manually reduced the setpoint by 1°C using manual operation 1. The linkage algorithm calculated at this time... The short-term trend is stable and insufficient to cross [a certain threshold]. k 2. kThe 35°C dead zone remained inactive. After a period of time following the shift to load reduction, due to… t 2. Rapid rise (see appendix) Figure 7 The linkage algorithm extracts this trend through positive fixed-value cruise and responds promptly. The speed was reduced (cruise 2), which is basically consistent with the operator's manual operation 3. Due to After reaching its peak, it quickly drops, and the linkage control algorithm returns to the set value (cruise 2). The descent was too rapid, triggering Cruise 3. Manual Operation 4, besides restoring the operational complexity of Manual Operation 3, also increased the setpoint, which can be considered equivalent to the sum of Cruise 2 and Cruise 3. Finally, The significant deviation from the setpoint indicates an issue with the operator's water-fuel matching and energy allocation, a long-term trend. Therefore, both Cruise 4 and Manual 5 reduced their setpoints. Based on this, the above summary is shown in Table 6 below: Table 6: Main Steam Temperature Setpoint Cruise Summary

[0078] This application embodiment is applied to a 1000MW ultra-supercritical unit. The main design parameters of the steam temperature section are shown in Table 7 below. This unit is an ultra-supercritical single-pass reheat once-through boiler, using Xiwang coal and Weiqiang coal as fuel. It employs a ball mill, intermediate storage pulverization, hot air delivery, and opposed combustion at the front and rear walls. Since the intermediate point temperature (superheat) setpoint has a significant impact on the water-to-coal ratio, and the water-to-coal ratio is coupled with the main steam pressure control, this embodiment of the application does not involve the intermediate point temperature during setpoint cruise, but selects a 1-hour application scenario during summer daytime when the unit consumes new energy.

[0079] Table 7: Design parameters of main steam temperature under typical operating conditions for a 1000MW unit

[0080] Appendix Figure 8 and attached Figure 9 The image shows the actual effect. See attached diagram. Figure 8 and attached Figure 9 As shown, the unit was in steady state from 0-20 minutes, with the load stable at 707±3MW, the main steam temperature setpoint at 600℃, and the secondary outlet steam temperature setpoint at 545℃. Because the secondary reducer valve in Section 2.1 fine-tunes the main steam temperature, its cruise dead zone... k 2. k The values ​​of 3 are relatively small, and due to the frequent short-term changes in the secondary outlet steam temperature, the setpoints are adjusted more frequently; the primary reduction valve fine-tunes the main steam temperature, and its setpoint cruise characteristic signals all show long-term trends, with limited cruise dead zones. k 2. kAll three values ​​are relatively high. Throughout the entire load change process, the main steam temperature and the secondary outlet steam temperature achieved constant value cruise, and the two-stage desuperheating water was more coordinated, requiring no manual intervention from the operators throughout the process.

[0081] This application addresses the problem of frequent setpoint adjustments required during deep peak shaving and rapid load changes in coal-fired power units under new power systems. It proposes an automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature. This method uses variational mode decomposition to decompose steam temperature characteristics and extract ground-state and dynamic components. Reverse and forward setpoint cruises are then used to achieve heat distribution rebalancing and proactive response, improving control quality. In this embodiment, the variational mode decomposition method extracts features from the steam-water system output. Based on the combination of these features, forward and reverse cruises of the main steam temperature and the secondary superheater outlet steam temperature are performed. Simultaneously, adjustments to the intermediate point temperature setpoint are suggested, achieving three-level linkage control of the main steam temperature from the water-fuel ratio to the primary and secondary desuperheaters. Finally, simulation and engineering applications verify the effectiveness of the method. Simulations show that this linkage control system has reasonable setpoint cruise timing and accurate direction judgment, and can partially replace or reduce manual intervention by operators. Application in a 1000MW ultra-supercritical unit shows that this method can adapt to the complex and ever-changing environment on site. During the load change process of peak shaving and frequency regulation, it can realize automatic cruise of set values ​​and no intervention by operators throughout the process, effectively reducing labor intensity and showing good engineering application prospects.

[0082] See attached document Figure 10 As shown, this application also discloses an automatic cruise control system for thermal power units based on multi-level linkage of main steam temperature, including: a model building module 1010, a simulation result output module 1020, an intrinsic mode generation module 1030, a modal component extraction module 1040, a cruise calculation module 1050, and an adjustment command generation module 1060.

[0083] For example, the model building module 1010 is used to abstract the main steam temperature system into a controlled object model, and adopts a first-order inertial plus time delay transfer function as the mathematical structure of the controlled object model.

[0084] For example, the simulation result output module 1020 is used to determine the characteristic parameters of the controlled object model through parameter identification and output the simulation results of the main steam temperature dynamic characteristics.

[0085] For example, the intrinsic mode generation module 1030 is used to take the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal as input, and process the input using variational mode decomposition to obtain multiple intrinsic mode components.

[0086] For example, the modal component extraction module 1040 is used to extract the ground state component representing the long-term change trend and the dynamic component representing the short-term change trend from the intrinsic modal component.

[0087] For example, the cruise calculation module 1050 is used to perform forward cruise calculation and reverse cruise calculation with the ground state component and the dynamic component as core inputs, and output the calculation results of forward cruise and reverse cruise.

[0088] For example, the adjustment instruction generation module 1060 is used to smooth and limit the calculation results of the forward cruise and the reverse cruise through an activation function, and output an automatic adjustment instruction.

[0089] Additionally, refer to the appendix Figure 11 As shown, a linkage control system based on constant-value cruise is illustrated, demonstrating the two-stage desuperheating water linkage control system designed in this application. The dashed boxes represent the combined... Figure 1 The original system, obtained from equations (3) and (4), directly connects the traditional two-stage desuperheating water control loop. Compared to the original system, the linkage control system performs variational mode decomposition feature extraction after the main steam temperature system outputs, performs combined calculations on the extracted features, and outputs the cruise adjustment amount of the setpoint to the operator's manual setpoint, thus realizing... and The closed-loop automatic cruise control system, however, only provides open-loop adjustment suggestions because the intermediate point temperature involves the water-to-coal ratio and is coupled with the main steam pressure control. In contrast, the original system's two control loops were independent and decoupled, requiring operators to manually adjust them continuously based on the output of the dashed box and the current unit operating conditions after setting the setpoint. This relied heavily on experience and increased workload. The linkage control system differs from the original system in that almost all flue gas flow disturbances are controlled by... Overcoming steam flow disturbances is almost entirely due to To overcome disturbances, the linkage control system uses setpoint cruise to integrate disturbance mitigation into steam temperature control. This achieves coarse adjustment of the water-to-coal ratio. Fine-tuning, Fine-tuned linkage to overcome.

[0090] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of this application, and this application is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this application, and these modifications and improvements are also considered to be within the scope of protection of this application.

Claims

1. An automatic cruise control method for thermal power units based on multi-level linkage of main steam temperature, characterized in that, include: The main steam temperature system is abstracted into a controlled object model, and a first-order inertial plus time delay transfer function is used as the mathematical structure of the controlled object model. The characteristic parameters of the controlled object model are determined by parameter identification, and the simulation results of the main steam temperature dynamic characteristics are output. The simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal are used as inputs, and the inputs are processed by variational mode decomposition to obtain multiple intrinsic mode components. The ground state component, which represents the long-term trend, and the dynamic component, which represents the short-term trend, are extracted from the intrinsic modal components. Using the ground state component and the dynamic component as core inputs, forward cruise calculation and reverse cruise calculation are performed, and the calculation results of forward cruise and reverse cruise are output. The calculation results of the forward cruise and the reverse cruise are smoothed and limited by the activation function, and an automatic adjustment command is output.

2. The method according to claim 1, characterized in that, The method further includes: Water-coal ratio adjustment is used as a coarse adjustment method to respond to adjustment commands for intermediate point temperature setpoints; The primary desuperheating water regulation is used as a fine-tuning method to respond to the adjustment command of the secondary superheater outlet steam temperature setpoint; The system uses secondary desuperheating water regulation as a fine-tuning method to respond to adjustment commands for the main steam temperature setpoint.

3. The method according to claim 1, wherein the transfer function model includes three undetermined parameters: gain, time constant, and lag time, characterized in that, The process of determining the characteristic parameters of the controlled object model through parameter identification includes: A step disturbance signal is applied to the desuperheating valve, and the real-time response data of the main steam temperature is recorded; The particle swarm optimization algorithm is used to iteratively optimize the gain, the time constant, and the lag time parameter until the fit between the output of the controlled object model and the real-time response data of the main steam temperature meets a preset threshold.

4. The method according to claim 1, characterized in that, The simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal are used as inputs, and the inputs are processed by variational mode decomposition to obtain multiple intrinsic mode components, including: The preset number of modal decompositions and the range of values ​​for the penalty parameter are defined. Using the minimum envelope entropy as the fitness function, an optimization algorithm is used to determine the target combination of the number of mode decompositions K and the penalty parameters. The input signal is decomposed using the target combination to obtain multiple intrinsic mode components; The number of modal decompositions ranges from 3 to 8, and the penalty parameter ranges from 100 to 3000.

5. The method according to claim 4, characterized in that, The extraction of the ground-state component representing the long-term trend and the dynamic component representing the short-term trend from the intrinsic modal components includes: The intrinsic mode component with the lowest frequency is identified as the ground state component; The intrinsic modal components whose frequency matches the operating cycle of the desuperheating valve are identified as the dynamic components; The intrinsic mode components at other frequencies are identified as noise signals and discarded.

6. The method according to claim 1, characterized in that, The calculation for performing forward cruise includes: Based on the dynamic component of the intermediate point temperature, the first adjustment amount of the secondary superheater outlet steam temperature setpoint is calculated and obtained; The first adjustment amount is combined with the dynamic component of the secondary superheater outlet steam temperature to calculate the second adjustment amount of the main steam temperature setpoint.

7. The method according to any one of claims 6, characterized in that, The reverse cruise calculation includes: Calculate the deviation between the ground state component and the main steam temperature setpoint, and calculate the deviation between the actual opening degree of the desuperheating valve and the target opening degree; The deviation between the ground state component and the main steam temperature setpoint, as well as the deviation between the actual opening degree and the target opening degree of the desuperheating valve, are fused and calculated using a preset conversion function to obtain the third adjustment amount of the secondary superheater outlet steam temperature setpoint. Based on the third adjustment amount and the deviation between the actual opening degree and the target opening degree of the desuperheating valve, an intermediate adjustment amount for the intermediate point temperature setpoint is generated.

8. The method according to claim 1, characterized in that, The activation function is an improved hyperbolic tangent function used to simulate the operational inertia of operators; wherein, the mathematical expression of the activation function includes: the output automatic adjustment command is zero when it is in the dead zone, the output automatic adjustment command is a constant value when it is in the saturation zone, and the output automatic adjustment command exhibits an S-shaped curve when it is in the linear zone.

9. The method according to claim 8, characterized in that, The parameters of the activation function include the dead zone and the saturation zone, wherein the adjustment of the dead zone range and the value of the saturation zone includes: The dead zone range is adjusted by defining the upper limit of the positive cutoff region and the lower limit of the negative cutoff region; By defining a positive saturation upper limit and a negative saturation lower limit, the constant value of the saturation region is adjusted.

10. An automatic cruise control system for thermal power units based on multi-level linkage of main steam temperature, characterized in that, include: The model building module is used to abstract the main steam temperature system into a controlled object model, and adopts a first-order inertial plus time delay transfer function as the mathematical structure of the controlled object model. The simulation result output module is used to determine the characteristic parameters of the controlled object model through parameter identification and output the simulation results of the main steam temperature dynamic characteristics. The intrinsic mode generation module is used to take the simulation results of the main steam temperature dynamic characteristics and the actual measured steam temperature signal as input, and process the input using variational mode decomposition to obtain multiple intrinsic mode components. The modal component extraction module is used to extract the ground state component, which represents the long-term trend, and the dynamic component, which represents the short-term trend, from the intrinsic modal components. The cruise calculation module is used to perform forward cruise calculation and reverse cruise calculation with the ground state component and the dynamic component as core inputs, and output the calculation results of forward cruise and reverse cruise. The adjustment instruction generation module is used to smooth and limit the calculation results of the forward cruise and the reverse cruise through an activation function, and output an automatic adjustment instruction.

Images