A thermal power unit steam temperature full working condition multi-model undisturbed switching control method and system

Abstract

This invention proposes a multi-model non-disruptive switching control method and system for steam temperature under all operating conditions in thermal power units, relating to the field of thermal power generation. The method includes: acquiring steam temperature data and load conditions in the steam temperature loop of the thermal power unit; constructing steam temperature models under different load conditions across all operating conditions based on adaptive weight positive and negative feedback using a slime mold algorithm, using the steam temperature data and load conditions as input; constructing a continuous system model set from the steam temperature models under different load conditions across all operating conditions, and then discretizing it into a discrete system model set; using the discrete system model set as input, controlling the steam temperature data under a single operating condition using a step-wise generalized prediction method; non-linearly fitting weight allocation parameters to multiple models across all operating conditions, and obtaining a flexible switching curve through smoothing; and controlling the steam temperature across all operating conditions using the flexible switching curve. This invention achieves non-disruptive switching between different operating conditions, realizing precise steam temperature control across the entire operating range of thermal power units with limited computing resources.

Description

Technical Field

[0001] This invention relates to the field of thermal power generation technology, specifically to a method and system for non-disruptive switching control of steam temperature under all operating conditions in thermal power units. Background Technology

[0002] The production process of thermal power units is complex, often characterized by strong coupling, nonlinear operating conditions, time-varying nature, multiple time scales, and numerous uncertainties. The steam temperature control loop exhibits significant time lag and nonlinear, time-varying characteristics, with its performance drifting dramatically with load changes. Traditional PID (Proportional-Integral-Derivative) control struggles to meet the demands of continuous operation, while conventional Model Predictive Control (MPC) relies on precise object models. Model mismatch becomes particularly pronounced when facing uncertainties such as load and coal quality fluctuations, and equipment aging. When model mismatch exceeds a certain level, the control performance of MPC significantly degrades. Furthermore, real-time control is highly sensitive to computational complexity; the computational complexity of conventional multi-model predictive control increases exponentially with the number of models, making it difficult to meet 1-second control cycle requirements. For thermal power units with large inertia, large delays, and nonlinearity, current methods cannot achieve continuous control of thermal power units with limited computational resources. Furthermore, while conventional segmented linearized multi-model architecture based on the thermal power production process can construct a set of local linear models under different operating conditions, the operating conditions of the unit change frequently and the boundaries are blurred during actual operation. Directly adopting hard switching strategies (such as abruptly changing model parameters or structures) can easily lead to abrupt changes in control quantities, system oscillations, or even instability. Summary of the Invention

[0003] This invention addresses the problems existing in the prior art by providing a method and system for seamless switching of steam temperature across multiple operating conditions in thermal power units, enabling seamless switching between different operating conditions.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] Acquire steam temperature data and load conditions in the steam temperature circuit of thermal power units; Based on steam temperature data and load conditions as input, an adaptive weight positive and negative feedback model of the slime mold algorithm is constructed for different load conditions across all operating conditions. The steam temperature model under different load conditions of the whole working condition is constructed into a continuous system model set and then discretized into a discrete system model set. Using a discrete system model set as input, a step-wise generalized prediction method is used to control steam temperature data under a single operating condition. Multiple models under all operating conditions are nonlinearly fitted with weighted parameters, and a flexible switching curve is obtained through smoothing. The temperature under all operating conditions is controlled by the flexible switching curve.

[0006] In some embodiments, the step of constructing steam temperature models under different load conditions for the entire operating condition based on adaptive weight positive and negative feedback of the slime mold algorithm, using steam temperature data and load conditions as input, includes: Construct a first-order inertial element model and obtain the parameters to be optimized; Optimal parameters are obtained through adaptive weight positive and negative feedback based on slime mold algorithm; The first-order inertial element model is used as the steam temperature model for thermal power units based on the optimal parameters. Parameter optimization was performed under different load conditions to obtain steam temperature models under different load conditions for the entire operating range.

[0007] In some embodiments, the step of acquiring steam temperature data and load conditions in the steam temperature circuit of a thermal power unit includes: Sensors are installed in the steam temperature circuit of a thermal power unit, which includes a desuperheater, a secondary desuperheater, and a secondary superheater, to perform cascade regulation control of steam temperature control, which includes two-stage regulation. For the first stage of regulation, the desuperheater spray flow rate is obtained, and the inlet and outlet temperature difference of the second stage desuperheater is obtained from the steam temperature data. For the second-stage regulation, the coal feed rate and the temperature difference between the inlet and outlet of the second-stage superheater in the steam temperature data are obtained. The current load conditions of thermal power units are collected synchronously.

[0008] In some embodiments, the process of constructing a first-order inertial element model and obtaining the parameters to be optimized includes: For the first stage of regulation, the desuperheater spray flow rate is used as the input and the inlet and outlet temperature difference of the second stage desuperheater is used as the output to construct a first-order inertial link model of the first stage. For the second-stage regulation, the coal feed rate is used as the input and the temperature difference between the inlet and outlet of the second-stage superheater is used as the output to construct a first-order inertial link model for the second stage. The first-order inertial element model and the second-order inertial element model are equivalent to a first-order inertial element model, and the transfer function corresponding to the first-order inertial element model is used as the parameter to be optimized.

[0009] In some embodiments, the process of obtaining optimal parameters through adaptive weight positive and negative feedback based on the slime mold algorithm includes: The actual values ​​of the parameters to be optimized are obtained by inputting steam temperature data and load information into the first-order inertial element model. Based on the slime mold algorithm, the predicted values ​​of the parameters to be optimized are obtained by simulating the position update equation with positive feedback and adjusting the adaptive weights with negative feedback. A fitness function is constructed using the predicted and actual values ​​of the parameters to be optimized. Based on the slime mold algorithm, the parameters of the first-order inertial link model are optimized by minimizing the fitness function through position updates and adaptive weight adjustments, and the optimal parameters are obtained.

[0010] In some embodiments, the process of using a discrete system model set as input, controlling steam temperature data under a single operating condition using a stepped generalized prediction method, nonlinearly fitting weight allocation parameters to multiple models for the entire operating condition, and obtaining a flexible switching curve through smoothing processing, and controlling the air temperature under the entire operating condition through the flexible switching curve includes: The optimal control quantity of the model at a certain moment under a single operating condition is calculated based on the step-type generalized predictive control algorithm, and the steam temperature data under the single operating condition is controlled based on the optimal control quantity. Based on steam temperature data under different single operating conditions, the weight allocation parameters of each model under all operating conditions are obtained by nonlinear fitting. When the operating conditions change, the weight allocation parameters of the model corresponding to the new operating conditions and the model corresponding to the old operating conditions are weighted and summed through soft switching. The summation value is smoothed according to the preset switching period and preset timer rules, and a flexible switching curve is output. The steam temperature is controlled under all operating conditions by using a flexible switching curve.

[0011] In some embodiments, the transfer function is formulated as follows: ; Where K is the gain and T is the time constant. Let s be the delay time, and s be the Laplace multiplier; The formula for the position update equation is as follows: ; ; in, X represents the current minimum fitness function value; X is the current position of the slime mold. and , where represents the positions of two randomly selected individuals from the population; t represents the current iteration number. Let p be the maximum number of iterations, and p be the threshold for escaping a local optimum. The first coefficient has a value in [- , Random oscillations between ] It is an oscillating variable; is the second coefficient, whose value is between [0,1] and eventually tends to 0; r is a random value in the interval [0,1]. For the fitness of slime mold X; The minimum fitness value across all iterations; The calculation formula is as follows: ; The formula for the adaptive weights is as follows: ; Where C represents The top half of the population; This represents the optimal fitness obtained in the current iteration. This represents the worst fitness obtained in the current iteration. This represents a fitness-increasing sequence; The formula for the fitness function is as follows: ; Where m is the number of samples. and represents the actual output value and the model output value, and k represents the sample.

[0012] This invention proposes a multi-model, disturbance-free switching control system for steam temperature of thermal power units under all operating conditions, comprising: The acquisition module is configured to acquire steam temperature data and load conditions in the steam temperature circuit of the thermal power unit. The full-condition module is configured to construct steam temperature models under different load conditions based on adaptive weight positive and negative feedback of the slime mold algorithm, using steam temperature data and load conditions as input. The discrete module is configured to construct a continuous system model set from the steam temperature model under different load conditions across the entire operating range, and then discretize it into a discrete system model set. The flexible switching module is configured to take a discrete system model set as input, use a step-wise generalized prediction method to control the steam temperature data under a single operating condition, nonlinearly fit the weight allocation parameters of multiple models under all operating conditions, and obtain the flexible switching curve through smoothing processing, and control the air temperature under all operating conditions through the flexible switching curve.

[0013] This invention proposes a computer device, comprising: At least one processor; and a memory storing a computer program that can run on the processor, wherein the processor executes the steps of the method for multi-model disturbance-free switching control of steam temperature under all operating conditions of a thermal power unit when executing the program.

[0014] This invention proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the described method for non-disruptive switching control of steam temperature under all operating conditions in thermal power units.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This invention proposes a multi-model non-disruptive switching control method and system for steam temperature under all operating conditions in thermal power units. The method includes: acquiring steam temperature data and load conditions in the steam temperature loop of the thermal power unit; constructing steam temperature models under different load conditions under all operating conditions based on adaptive weight positive and negative feedback using the slime mold algorithm, using the steam temperature data and load conditions as input; constructing a continuous system model set from the steam temperature models under different load conditions under all operating conditions, and discretizing it into a discrete system model set; using the discrete system model set as input, controlling the steam temperature data under a single operating condition using a step-wise generalized prediction method, nonlinearly fitting weight allocation parameters to multiple models under all operating conditions, and obtaining a flexible switching curve through smoothing processing, and controlling the steam temperature under all operating conditions through the flexible switching curve.

[0016] For the steam temperature loop of thermal power units with large inertia, large delay, and nonlinearity, this invention aims to study a robust control scheme suitable for industrial real-time scenarios with low computational requirements. By fitting the model transfer function parameters of multiple operating conditions under full operating conditions using a slime mold algorithm, and combining this with a stepped generalized predictive control algorithm, the online computation intensity is reduced to meet real-time control requirements. Furthermore, a dual-model soft-switching framework based on operating condition feature weighting is proposed to achieve bumpless switching between different operating conditions. The proposed multi-model stepped generalized predictive controller achieves precise steam temperature control across the entire operating range of thermal power units with limited computational resources, significantly reducing computational load, improving real-time performance, and achieving bumpless switching throughout the entire process.

[0017] This invention studies a stepped generalized predictive controller suitable for industrial real-time scenarios with relatively low computational requirements, as well as a robust multi-model flexible switching method, which constitutes a nonlinear multi-model stepped generalized predictive controller. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0019] Figure 1 The flowchart of a multi-model non-disruptive switching control method for steam temperature of thermal power units under all operating conditions provided by the present invention is shown.

[0020] Figure 2 This invention provides a module diagram of a multi-model, non-disruptive switching control system for steam temperature of thermal power units under all operating conditions.

[0021] Figure 3 A schematic diagram of the structure of an embodiment of the computer device provided by the present invention.

[0022] Figure 4 This is a schematic diagram of an embodiment of the computer-readable storage medium provided by the present invention.

[0023] Figure 5 This invention provides a schematic diagram of steam temperature cascade control for a multi-model, non-disruptive switching control method for steam temperature under all operating conditions in thermal power units.

[0024] Figure 6 The present invention provides a soft switching framework diagram for a multi-model non-disruptive switching control method for steam temperature of thermal power units under all operating conditions. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention. It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application.

[0026] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities or parameters with the same name but different names. It is clear that "first" and "second" are only for the convenience of expression and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0027] This invention proposes a multi-model, disturbance-free switching control method for steam temperature of thermal power units under all operating conditions. Please refer to [link / reference]. Figure 1 , Figure 5 and Figure 6 ,include: Acquire steam temperature data and load conditions in the steam temperature circuit of thermal power units; Based on steam temperature data and load conditions as input, an adaptive weight positive and negative feedback model of the slime mold algorithm is constructed for different load conditions across all operating conditions. The steam temperature model under different load conditions of the whole working condition is constructed into a continuous system model set and then discretized into a discrete system model set. Using a discrete system model set as input, a step-wise generalized prediction method is used to control steam temperature data under a single operating condition. Multiple models under all operating conditions are nonlinearly fitted with weighted parameters, and a flexible switching curve is obtained through smoothing. The temperature under all operating conditions is controlled by the flexible switching curve.

[0028] In this invention, the more specific steps are as follows: Step 1: Establish a steam temperature model for thermal power units based on the slime mold algorithm.

[0029] A schematic diagram of steam temperature control cascade regulation is shown below. Figure 5 As shown. In the first-stage regulation, the desuperheater spray water flow rate is used as the system input, and the temperature difference between the inlet and outlet temperatures of the second-stage desuperheater is used as the system output for control. In the second-stage regulation, the coal feed rate is used as the system input, and the temperature difference between the inlet and outlet temperatures of the second-stage superheater is used as the system output for control. The controlled object is selected as a first-order inertial element as the model, as shown in the formula: ; Where K is the gain and T is the time constant. The delay time, s, is the Laplace multiplier. The parameters of the transfer function are used as parameters to be optimized.

[0030] The slime mold algorithm is inspired by the dispersal and foraging behavior of slime molds. This model uses adaptive weights to achieve a positive and negative feedback process, ensuring rapid convergence while maintaining a certain perturbation rate to avoid getting trapped in local optima.

[0031] The food-approaching behavior of slime molds is simulated as a mathematical equation, and its position update equation is as follows (taking the minimum value as the optimal solution): ; ; ; In the formula: X represents the current minimum fitness function value; X is the current position of the slime mold. and , where represents the positions of two randomly selected individuals from the population; t represents the current iteration number. This represents the maximum number of iterations. The coefficient is [-]. , Random oscillations between ] It is an oscillating variable; is a coefficient whose value is between [0,1] and eventually tends to 0; r is a random value in the interval [0,1]. For the fitness of slime mold X; This is the minimum fitness value across all iterations.

[0032] The expression for the adaptive weight W of slime mold is: ; C represents The top half of the population. This represents the optimal fitness obtained in the current iteration. This represents the worst fitness obtained in the current iteration. This represents a fitness-increasing sequence.

[0033] The fitness function is constructed based on the deviation between the model and the actual output: ; Where m is the number of samples. and These are the actual output values ​​and the model output values.

[0034] Step 2: Based on Step 1, obtain the steam temperature model under different load conditions across the entire operating range, thereby constructing a continuous system model set (CSMS) for steam temperature under all operating conditions. For ease of control, the CSMS is discretized into a discrete system model set (DSMS). In the zero-order hold method, for the transfer function of the continuous system, its discretized transfer function can be obtained by replacing it after discretization of the sampling period.

[0035] Step 3: Control the steam temperature of the steam temperature model under a single working condition using a stepwise generalized prediction method, and perform nonlinear fitting on the weight allocation parameters of the multi-model under all working conditions. Smoothly fuse the outputs of multiple models under adjacent working conditions to achieve flexible switching between different working conditions.

[0036] The optimal control quantity η at time t in the step-wise generalized predictive control algorithm is calculated as follows:

[0037] In the formula, To control the weights. w is the expected value of the system. It is a step decay factor. This is the transpose of the step response matrix. This is the final attenuation constant.

[0038]

[0039] In the formula, f is the free response vector. This is the first free response quantity. This is the second free response quantity;

[0040] The step response matrix, The final step response constant of the system; The first step response coefficient, The step response coefficient of a final-position system; The second step response coefficient, is the step response coefficient of the two-terminal system.

[0041] The formula for calculating W is:

[0042] In the formula, j is the first variation coefficient, u(k) is the system input, v(k) is the system feedforward, and g is the system step response coefficient. This is the current actual output of the object; This is the actual set value; To set the softening factor, . Let j be the softening factor.

[0043] Weights are assigned to multiple sub-models or sub-control variables, and then weighted summation is performed to obtain the global model or global control variable, using a soft-switching approach. This method uses the calculated weights to sum the values ​​of each sub-controller to obtain the final controller output.

[0044] All sub-models in the discrete system model set can be selected and loaded into the current model and target model regions, based on the operating condition characteristic parameters, corresponding to... Figure 6 In the program, the old model TM1sel and the new model TM2Sel are used. When the operating conditions change, TM2Sel selects the model corresponding to the new operating conditions (referred to as the "new model") and loads it into the target model area, while TM1Sel retains the model corresponding to the original operating conditions (referred to as the "old model"). The program synchronously calculates the control law output for the old and new models and performs soft switching calculations according to the following rules: ; in, Output for the old model. Output for the new model.

[0045] In the formula, TP is the model soft switching period, which can be set manually; TT is the model switching timer, representing the time elapsed for model switching. After the switching is completed, TM1Sel follows TM2Sel to select the new model. This completes one model soft switching process.

[0046] This invention employs a step-by-step generalized predictive control method, which transforms the complex matrix inversion operation in traditional generalized predictive control into a low-order algebraic operation. By introducing step-by-step control incremental planning, the complexity of solving the control law is reduced from O(n³) to O(n), meeting the requirements of second-level real-time control.

[0047] This invention is based on a dual-model soft handover framework weighted by operating condition characteristics. It eliminates disturbances during model switching and achieves seamless handover across all operating conditions.

[0048] This invention addresses the need for real-time control cycles on the order of seconds in the steam temperature circuit of thermal power units, thus proposing a stepped generalized predictive control method. To address the issues of sudden changes in control inputs, system oscillations, and even instability that can occur during hard switching between different operating conditions across the entire operating range, a nonlinear multi-model switching technique is proposed. This technique uses a dynamic weight allocation mechanism to smoothly fuse the outputs of multiple models for adjacent operating conditions.

[0049] In some embodiments, the step of constructing steam temperature models under different load conditions for the entire operating condition based on adaptive weight positive and negative feedback of the slime mold algorithm, using steam temperature data and load conditions as input, includes: Construct a first-order inertial element model and obtain the parameters to be optimized; Optimal parameters are obtained through adaptive weight positive and negative feedback based on slime mold algorithm; The first-order inertial element model is used as the steam temperature model for thermal power units based on the optimal parameters. Parameter optimization was performed under different load conditions to obtain steam temperature models under different load conditions for the entire operating range.

[0050] The slime mold algorithm is used to determine the adaptive weights of different sub-models (corresponding to different load conditions) in the full-condition model. Based on the diffusion and foraging behavior of slime molds, their food-approaching behavior is simulated to obtain the position update equation and the adaptive weights of the slime molds. The output of the cascade control model can be used as feedback information to dynamically adjust these weights. When there is a large deviation between the steam temperature output by the cascade control model and the steam temperature predicted by the full-condition model, it indicates that the current weight allocation of the full-condition model may be unreasonable. At this time, the positive and negative feedback mechanism of the slime mold algorithm is used to adjust the weights of different sub-models according to the magnitude and direction of the deviation, with the goal of minimizing the fitness function, to optimize the parameters of the first-order inertial link model of steam temperature cascade control.

[0051] In some embodiments, please refer to Figure 5 The steps for obtaining steam temperature data and load conditions in the steam temperature circuit of the thermal power unit include: Sensors are installed in the steam temperature circuit of a thermal power unit, which includes a desuperheater, a secondary desuperheater, and a secondary superheater, to perform cascade regulation control of steam temperature control, which includes two-stage regulation. For the first stage of regulation, the desuperheater spray flow rate is obtained, and the inlet and outlet temperature difference of the second stage desuperheater is obtained from the steam temperature data. For the second-stage regulation, the coal feed rate and the temperature difference between the inlet and outlet of the second-stage superheater in the steam temperature data are obtained. The current load conditions of thermal power units are collected synchronously.

[0052] In the steam temperature circuit of the thermal power unit, appropriate sensors are installed to acquire relevant data. For the first stage of regulation, the desuperheater spray water flow rate is collected as the system input signal, and the inlet and outlet temperatures of the second-stage desuperheater are also collected. For the second stage of regulation, the coal feed rate is collected as the system input signal, and the inlet and outlet temperatures of the second-stage superheater are also collected.

[0053] The collected temperature, flow rate, and other data are filtered to remove noise interference and improve data quality. Mean filtering or median filtering algorithms are used, and an appropriate filter window size is selected based on data characteristics and actual needs. The data is then normalized to unify data with different dimensions into the [0,1] interval.

[0054] In some embodiments, please refer to Figure 5 The process of constructing a first-order inertial element model and obtaining the parameters to be optimized includes: For the first stage of regulation, the desuperheater spray flow rate is used as the input and the inlet and outlet temperature difference of the second stage desuperheater is used as the output to construct a first-order inertial link model of the first stage. For the second-stage regulation, the coal feed rate is used as the input and the temperature difference between the inlet and outlet of the second-stage superheater is used as the output to construct a first-order inertial link model for the second stage. The first-order inertial element model and the second-order inertial element model are equivalent to a first-order inertial element model, and the transfer function corresponding to the first-order inertial element model is used as the parameter to be optimized.

[0055] A first-order inertial element is selected as the model structure based on the controlled object. For the first-stage regulation, the desuperheater spray flow rate is used as the input, and the temperature difference between the inlet and outlet temperatures of the second-stage desuperheater is used as the output to construct a first-stage first-order inertial element model. For the second-stage regulation, the coal feed rate is used as the input, and the temperature difference between the inlet and outlet temperatures of the second-stage superheater is used as the output to construct a second-stage first-order inertial element model.

[0056] In a steam temperature control cascade system, the output of the first-stage regulator (the outlet temperature of the second-stage desuperheater) affects the input of the second-stage regulator (the inlet temperature of the second-stage superheater). In other words, changes in the temperature difference of the first-stage regulator alter the steam temperature entering the second-stage superheater, thus affecting the temperature difference response of the second-stage regulator. When two first-order inertial elements are connected in series, it forms a second-order system, but with a relatively small sampling period, it can be approximated as a series-connected first-order inertial element model.

[0057] Further experiments were conducted to verify the accuracy of the equivalent first-order inertial element model. Under different input signals, the actual output of the equivalent model was compared with the output of the actual cascade control system. If the two match within the allowable error range, the equivalent model is considered reasonable. If the error is large, the model needs to be corrected. The inputs of the equivalent first-order inertial element model for cascade control are controlled variables such as desuperheater spray flow rate and coal feed rate, and the outputs are steam temperature-related parameters.

[0058] In some embodiments, please refer to Figure 6 The process of obtaining optimal parameters through adaptive weight positive and negative feedback based on the slime mold algorithm includes: The actual values ​​of the parameters to be optimized are obtained by inputting steam temperature data and load information into the first-order inertial element model. Based on the slime mold algorithm, the predicted values ​​of the parameters to be optimized are obtained by simulating the position update equation with positive feedback and adjusting the adaptive weights with negative feedback. A fitness function is constructed using the predicted and actual values ​​of the parameters to be optimized. Based on the slime mold algorithm, the parameters of the first-order inertial link model are optimized by minimizing the fitness function through position updates and adaptive weight adjustments, and the optimal parameters are obtained.

[0059] In positive feedback simulation, when the slime mold algorithm simulates the position update equation, it updates the position in potentially better directions based on the currently found better solution, similar to the behavior of slime molds gathering towards food sources, thereby obtaining predicted values ​​for the parameters to be optimized. If the predicted parameter value in a certain direction makes the model output closer to the expected value, the algorithm will increase the exploration effort in that direction.

[0060] Simultaneously, an adaptive weighting mechanism comes into play. When the predicted values ​​of certain parameters cause the model output to deviate significantly from the expected values, the adaptive weights reduce the influence of these parameters in subsequent updates, acting as a negative feedback correction. If the predicted value of a parameter causes the model output to be too high, the algorithm will reduce the weight of that parameter in subsequent position updates.

[0061] The fitness function is constructed using the predicted and actual values ​​of the parameters to be optimized. The design of the fitness function reflects the evaluation criteria for positive and negative feedback. If the predicted and actual values ​​are close, it indicates good performance, and the fitness function value will be small, aiming to minimize it, which is a positive evaluation of positive feedback. Conversely, if the difference between the two is large, the fitness function value will be large, indicating that the model needs adjustment, which is a warning signal of negative feedback.

[0062] Based on the slime mold algorithm, the goal is to minimize the fitness function by continuously exploring better parameter combinations through position updates. When a parameter combination that significantly reduces the fitness function value is found, the algorithm continues to search in that direction, which is a positive feedback-driven optimization process.

[0063] The negative feedback mechanism is always present during position updates and adaptive weight adjustments. If a new parameter combination causes the fitness function value to increase instead of decrease, it indicates that the optimization direction may be incorrect. The algorithm will correct this through adaptive weight adjustments and position updates to avoid continuing in the wrong direction and ensure that the optimization process proceeds towards the global optimum. Ultimately, the optimal parameters of the first-order inertial link model are obtained.

[0064] In some embodiments, please refer to Figure 6 The process of using a discrete system model set as input, controlling steam temperature data under a single operating condition using a stepped generalized prediction method, nonlinearly fitting weight allocation parameters to multiple models for all operating conditions, and obtaining a flexible switching curve through smoothing, and controlling the air temperature under all operating conditions through the flexible switching curve includes: The optimal control quantity of the model at a certain moment under a single operating condition is calculated based on the step-type generalized predictive control algorithm, and the steam temperature data under the single operating condition is controlled based on the optimal control quantity. Based on steam temperature data under different single operating conditions, the weight allocation parameters of each model under all operating conditions are obtained by nonlinear fitting. When the operating conditions change, the weight allocation parameters of the model corresponding to the new operating conditions and the model corresponding to the old operating conditions are weighted and summed through soft switching. The summation value is smoothed according to the preset switching period and preset timer rules, and a flexible switching curve is output. The steam temperature is controlled under all operating conditions by using a flexible switching curve.

[0065] Based on the step-wise generalized predictive control algorithm, the optimal control quantity at time k is calculated, and the steam temperature is controlled under a single operating condition based on the calculated optimal control quantity.

[0066] Based on steam temperature data and model characteristics under different operating conditions, nonlinear fitting is performed on the weight allocation parameters of the multi-model under all operating conditions. The performance of each sub-model under different operating conditions is analyzed to determine an appropriate weight allocation method so that the weights can accurately reflect the importance of each sub-model under different operating conditions.

[0067] The sub-models corresponding to the old and new operating conditions are the old model and the new model, respectively. Using a soft handover method, the sub-controllers of the old and new models for adjacent operating conditions are weighted and summed using calculated weights. Soft handover calculations are performed based on the model soft handover cycle and the model handover timer to achieve smooth fusion of the outputs of multiple models for adjacent operating conditions.

[0068] A flexible switching curve is generated based on the output data of the smoothed and fused model. This curve reflects the transition process of steam temperature control under different operating conditions. When the operating conditions change, the model corresponding to the new operating condition is selected from the discrete system model set according to the characteristic parameters of the operating condition, and the flexible switching curve is used to achieve continuous and stable control of the air temperature under all operating conditions.

[0069] In some embodiments, the transfer function is formulated as follows: ; Where K is the gain and T is the time constant. Let s be the delay time, and s be the Laplace multiplier; The formula for the position update equation is as follows: ; ; in, X represents the current minimum fitness function value; X is the current position of the slime mold. and , where represents the positions of two randomly selected individuals from the population; t represents the current iteration number. Let p be the maximum number of iterations, and p be the threshold for escaping a local optimum. The first coefficient has a value in [- , Random oscillations between ] It is an oscillating variable; is the second coefficient, whose value is between [0,1] and eventually tends to 0; r is a random value in the interval [0,1]. For the fitness of slime mold X; The minimum fitness value across all iterations; The calculation formula is as follows: ; The formula for the adaptive weights is as follows: ; Where C represents The top half of the population; This represents the optimal fitness obtained in the current iteration. This represents the worst fitness obtained in the current iteration. This represents a fitness-increasing sequence; The formula for the fitness function is as follows: ; Where m is the number of samples. and represents the actual output value and the model output value, and k represents the sample.

[0070] This invention proposes a multi-model, disturbance-free switching control system for steam temperature in thermal power units under all operating conditions. Please refer to [link / reference]. Figure 2 ,include: The acquisition module is configured to acquire steam temperature data and load conditions in the steam temperature circuit of the thermal power unit. The full-condition module is configured to construct steam temperature models under different load conditions based on adaptive weight positive and negative feedback of the slime mold algorithm, using steam temperature data and load conditions as input. The discrete module is configured to construct a continuous system model set from the steam temperature model under different load conditions across the entire operating range, and then discretize it into a discrete system model set. The flexible switching module is configured to take a discrete system model set as input, use a step-wise generalized prediction method to control the steam temperature data under a single operating condition, nonlinearly fit the weight allocation parameters of multiple models under all operating conditions, and obtain the flexible switching curve through smoothing processing, and control the air temperature under all operating conditions through the flexible switching curve.

[0071] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 3 As shown, an embodiment of the present invention also provides a computer device 30, which includes a processor 310 and a memory 320. The memory 320 stores a computer program 321 that can be run on the processor. When the processor 310 executes the program, it performs the steps of the method described above.

[0072] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 4 As shown, embodiments of the present invention also provide a computer-readable storage medium 40, which stores a computer program 410 that, when executed by a processor, performs the methods described above.

[0073] Embodiments of the present invention may also include a corresponding computer device. The computer device includes a memory, at least one processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes any of the methods described above when executing the program.

[0074] The memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as program instructions / modules in the embodiments of this application. The processor executes various functional applications and data processing of the device by running the non-volatile software programs, instructions, and modules stored in the memory, thereby implementing the above-described method.

[0075] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the device. Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0076] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The storage medium for the program can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The above computer program embodiments can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0077] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the system as a whole. Those skilled in the art can implement the functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0078] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. The sequence numbers of the disclosed embodiments of this invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0079] It should be understood that, as used herein, the singular form “a” is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, “and / or” refers to any and all possible combinations of one or more of the associated listed items.

[0080] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A method for non-disruptive switching control of steam temperature in thermal power units under multiple operating conditions using multiple models, characterized in that, include: Acquire steam temperature data and load conditions in the steam temperature circuit of thermal power units; Based on steam temperature data and load conditions as input, an adaptive weight positive and negative feedback model of the slime mold algorithm is constructed for different load conditions across all operating conditions. The steam temperature model under different load conditions of the whole working condition is constructed into a continuous system model set and then discretized into a discrete system model set. Using a discrete system model set as input, a step-wise generalized prediction method is used to control steam temperature data under a single operating condition. Multiple models under all operating conditions are nonlinearly fitted with weighted parameters, and a flexible switching curve is obtained through smoothing. The temperature under all operating conditions is controlled by the flexible switching curve.

2. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 1, characterized in that, The steps of constructing steam temperature models under different load conditions for the entire operating condition based on adaptive weight positive and negative feedback using the slime mold algorithm, with steam temperature data and load conditions as input, include: Construct a first-order inertial element model and obtain the parameters to be optimized; Optimal parameters are obtained through adaptive weight positive and negative feedback based on slime mold algorithm; The first-order inertial element model is used as the steam temperature model for thermal power units based on the optimal parameters. Parameter optimization was performed under different load conditions to obtain steam temperature models under different load conditions for the entire operating range.

3. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 2, characterized in that, The steps for obtaining steam temperature data and load conditions in the steam temperature circuit of a thermal power unit include: Sensors are installed in the steam temperature circuit of a thermal power unit, which includes a desuperheater, a secondary desuperheater, and a secondary superheater, to perform cascade regulation control of steam temperature control, which includes two-stage regulation. For the first stage of regulation, the desuperheater spray flow rate is obtained, and the inlet and outlet temperature difference of the second stage desuperheater is obtained from the steam temperature data. For the second-stage regulation, the coal feed rate and the temperature difference between the inlet and outlet of the second-stage superheater in the steam temperature data are obtained. The current load conditions of thermal power units are collected synchronously.

4. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 3, characterized in that, The process of constructing a first-order inertial element model and obtaining the parameters to be optimized includes: For the first stage of regulation, the desuperheater spray flow rate is used as the input and the inlet and outlet temperature difference of the second stage desuperheater is used as the output to construct a first-order inertial link model of the first stage. For the second-stage regulation, the coal feed rate is used as the input and the temperature difference between the inlet and outlet of the second-stage superheater is used as the output to construct a first-order inertial link model for the second stage. The first-order inertial element model and the second-order inertial element model are equivalent to a first-order inertial element model, and the transfer function corresponding to the first-order inertial element model is used as the parameter to be optimized.

5. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 4, characterized in that, The process of obtaining optimal parameters through adaptive weight positive and negative feedback based on the slime mold algorithm includes: The actual values ​​of the parameters to be optimized are obtained by inputting steam temperature data and load information into the first-order inertial element model. Based on the slime mold algorithm, the predicted values ​​of the parameters to be optimized are obtained by simulating the position update equation with positive feedback and adjusting the adaptive weights with negative feedback. A fitness function is constructed using the predicted and actual values ​​of the parameters to be optimized. Based on the slime mold algorithm, the parameters of the first-order inertial link model are optimized by minimizing the fitness function through position updates and adaptive weight adjustments, and the optimal parameters are obtained.

6. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 1, characterized in that, The process of using a discrete system model set as input, controlling steam temperature data under a single operating condition using a stepped generalized prediction method, nonlinearly fitting weighted parameters to multiple models for all operating conditions, and obtaining a flexible switching curve through smoothing, and controlling the air temperature under all operating conditions through the flexible switching curve includes: The optimal control quantity of the model at a certain moment under a single operating condition is calculated based on the step-type generalized predictive control algorithm, and the steam temperature data under the single operating condition is controlled based on the optimal control quantity. Based on steam temperature data under different single operating conditions, the weight allocation parameters of each model under all operating conditions are obtained through nonlinear fitting. When the operating conditions change, the weight allocation parameters of the model corresponding to the new operating conditions and the model corresponding to the old operating conditions are weighted and summed through soft switching. The summation value is smoothed according to the preset switching period and preset timer rules, and a flexible switching curve is output. The steam temperature is controlled under all operating conditions by using a flexible switching curve.

7. The method for non-disruptive switching control of steam temperature in thermal power units under all operating conditions according to claim 5, characterized in that, The formula for the transfer function is as follows: ； Where K is the gain and T is the time constant. Let s be the delay time, and s be the Laplace multiplier; The formula for the position update equation is as follows: ； ； in, X represents the current minimum fitness function value; X is the current position of the slime mold. and , where represents the positions of two randomly selected individuals from the population; t represents the current iteration number. Let p be the maximum number of iterations, and p be the threshold for escaping a local optimum. The first coefficient has a value in [- , Random oscillations between ] It is an oscillating variable; is the second coefficient, whose value is between [0,1] and eventually tends to 0; r is a random value in the interval [0,1]. For the fitness of slime mold X; The minimum fitness value across all iterations; The calculation formula is as follows: ； The formula for the adaptive weights is as follows: ； Where C represents The top half of the population; This represents the optimal fitness obtained in the current iteration. This represents the worst fitness obtained in the current iteration. This represents a sequence with increasing fitness. The formula for the fitness function is as follows: ； Where m is the number of samples. and represents the actual output value and the model output value, and k represents the sample.

8. A multi-model, non-disruptive switching control system for steam temperature under all operating conditions in thermal power units, characterized in that, include: The acquisition module is configured to acquire steam temperature data and load conditions in the steam temperature circuit of the thermal power unit. The full-condition module is configured to construct steam temperature models under different load conditions based on adaptive weight positive and negative feedback of the slime mold algorithm, using steam temperature data and load conditions as input. The discrete module is configured to construct a continuous system model set from the steam temperature model under different load conditions across the entire operating range, and then discretize it into a discrete system model set. The flexible switching module is configured to take a discrete system model set as input, use a step-wise generalized prediction method to control the steam temperature data under a single operating condition, nonlinearly fit the weight allocation parameters of multiple models under all operating conditions, and obtain the flexible switching curve through smoothing processing, and control the air temperature under all operating conditions through the flexible switching curve.

9. A computer device, comprising: At least one processor; The processor includes a memory storing a computer program that can run on the processor, characterized in that the processor executes the program by performing the steps of the multi-model non-disruptive switching control method for steam temperature of a thermal power unit under all operating conditions as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it performs the steps of the multi-model non-disruptive switching control method for steam temperature of thermal power units under all operating conditions as described in any one of claims 1 to 7.

Images