Method and system for evaluating steam extraction regulation and control performance of rotary partition plate

Abstract

The invention discloses a method and system for evaluating the steam extraction regulation and control performance of a rotating partition plate, and relates to the technical field of evaluation of the steam extraction regulation and control performance of the rotating partition plate, and the method comprises the steps: collecting the real-time data of the steam extraction regulation and control operation of the rotating partition plate, and constructing a working condition-state-response data set through preprocessing; the method comprises the following steps: constructing a non-linear coupling response prediction model taking a rotating partition plate angle as a driving variable based on rotating partition plate steam extraction regulation and control operation real-time data, and constructing a three-layer performance index system through a deviation between the non-linear coupling response prediction model and an actual monitoring value; and constructing a regulation and control efficiency mapping model through the partition plate angle and the three-layer performance indexes, and classifying and identifying performance index vectors output by the regulation and control efficiency mapping model. According to the method, steam extraction performance accurate modeling and classified regulation and control based on rotating partition plate angle driving are achieved, the system response precision, the regulation stability and the strategy matching performance are improved, and the method adapts to the multi-working-condition dynamic operation requirement.

Description

Technical Field

[0001] The present invention relates to the technical field of rotary baffle steam extraction control performance evaluation, and in particular to a rotary baffle steam extraction control performance evaluation method and system. Background Art

[0002] With the continuous improvement of the intelligent operation level of power plant thermal systems, the control strategy of steam turbine extraction is becoming increasingly complex. As the key executive component for regulating the extraction flow and pressure, the operating stability and control response performance of the rotating diaphragm have a significant impact on the system's energy efficiency level, load matching capability and economic indicators. At present, power plants have generally deployed high-frequency sampling operation data acquisition systems, which can record multi-source operating data such as rotating diaphragm angle, extraction flow, pressure, main steam parameters, etc. in real time, providing basic data support for the performance evaluation of extraction control behavior. With the development of data-driven modeling and nonlinear system analysis theory, more and more studies have begun to try to construct dynamic control evaluation models from the perspective of mechanism-data fusion to achieve quantitative evaluation of the rotating diaphragm control process.

[0003] There are still many technical bottlenecks in the existing evaluation methods for the control performance of rotating diaphragms. Traditional evaluation methods are mostly based on single-point steady-state parameter threshold judgments or empirical criteria. They lack the ability to conduct coupling analysis from the perspective of the entire dynamic response process, and it is difficult to accurately capture the potential nonlinear disturbances in the response chain caused by changes in the angle of the rotating diaphragm. Most existing methods ignore the multiple modulation effects of operating variables on response parameters, and are unable to accurately predict the trend of changes in steam extraction response under actual operating conditions. There is a lack of quantitative models with strong versatility and high prediction accuracy. There is currently a lack of a systematic performance evaluation index system, which cannot comprehensively reflect the deviation characteristics, dynamic characteristics and steady-state stability. As a result, the control effectiveness analysis often remains at the level of empirical judgment and lacks a reusable and traceable evaluation structure. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing rotating baffle steam extraction control performance evaluation method has the problems of single evaluation dimension and insufficient dynamic process modeling capability, the inability to accurately predict the nonlinear behavior of the system response caused by angle changes, the lack of a quantifiable and traceable performance indicator system, and how to achieve full-dimensional identification and classification evaluation of control performance under complex working conditions.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for evaluating the performance of rotating baffle steam extraction control, comprising collecting real-time data of rotating baffle steam extraction control operation, and constructing a working condition-state-response data set through preprocessing; constructing a nonlinear coupling response prediction model with the rotating baffle angle as the driving variable based on the real-time data of rotating baffle steam extraction control operation, and constructing a three-layer performance index system through the deviation between the nonlinear coupling response prediction model and the actual monitoring value; constructing a control efficiency mapping model through the baffle angle and the three-layer performance index, and classifying and identifying the performance index vector output by the control efficiency mapping model; the three-layer performance index system includes deviation indicators, dynamic process indicators and steady-state stability indicators; the nonlinear coupling response prediction model includes introducing auxiliary working condition variables to define the predicted extraction response value, constructing a nonlinear coupling response prediction model, fitting the real-time data of rotating baffle steam extraction control operation during the construction process for dynamic estimation, and predicting the response quantity driven by the rotating baffle angle.

[0007] As an optimal solution of the rotary baffle steam extraction control performance evaluation method described in the present invention, the real-time data of the rotary baffle steam extraction control operation collected includes the baffle rotation angle, steam extraction pressure and flow, turbine main steam flow, load, speed, main steam temperature and steam extraction temperature.

[0008] As a preferred solution of the rotary baffle steam extraction control performance evaluation method described in the present invention, the pre-processed construction of the working condition-state-response data set includes time synchronization, missing completion and outlier elimination of the real-time data of the rotary baffle steam extraction control operation, and the use of moving average filtering and standardization processing to form a structured working condition-state-response data set.

[0009] As a preferred embodiment of the rotary diaphragm extraction control performance evaluation method of the present invention, the nonlinear coupling response prediction model with the rotary diaphragm angle as the driving variable is constructed based on the real-time data of the rotary diaphragm extraction control operation, including constructing a nonlinear coupling model with the rotary diaphragm angle as the main driving variable based on the collected rotary diaphragm angle θ, extraction pressure ζ, turbine load ψ, and main steam temperature η, expressed as:

[0010]

[0011] Where θ represents the current angle of the rotating diaphragm, α represents the target diaphragm angle setting value, β represents the angle adjustment sensitivity factor, ζ represents the current extraction steam pressure, γ represents the load disturbance adjustment coefficient, ψ represents the current turbine load, ρ represents the expected load value, κ represents the temperature disturbance correction factor, η represents the main steam temperature, and v represents the steady-state temperature cycle parameter.

[0012] As a preferred solution of the rotating baffle steam extraction control performance evaluation method described in the present invention, the three-layer performance index system is constructed through the deviation between the nonlinear coupling response prediction model and the actual monitoring value, including, based on the deviation between the output value of the nonlinear coupling model and the actual monitoring value, constructing a three-layer performance index system consisting of execution deviation indicators, dynamic process indicators and steady-state stability indicators, which are used to characterize the steam extraction accuracy, control response behavior and system fluctuation characteristics respectively.

[0013] As a preferred solution of the rotary baffle steam extraction control performance evaluation method described in the present invention, the control efficiency mapping model is constructed through the baffle angle and three-layer performance indicators, including determining the input factor vector based on the collected real-time data of the rotary baffle steam extraction control operation and the variables actually used in the nonlinear coupling response prediction model; using the principal component analysis method to transform the input factors, extracting the characteristic factor combination with the maximum explained variance, outputting the principal factor tensor, and constructing the output mapping relationship of the control efficiency mapping model.

[0014] As a preferred solution of the rotating diaphragm steam extraction control performance evaluation method described in the present invention, the classification and identification of the performance index vector output by the control efficiency mapping model includes classification and identification based on the mapping relationship output by the control efficiency mapping model, and dividing it into four performance states: stability priority type, response acceleration type, accuracy compensation type and comprehensive weakening type, and outputting the corresponding rotating diaphragm control strategy parameter vector for each performance type.

[0015] Another object of the present invention is to provide a rotating baffle steam extraction control performance evaluation system, which can construct a control efficiency mapping model through the baffle angle and three-layer performance indicators, and classify and identify the performance indicator vectors output by the control efficiency mapping model, thereby solving the problem that the current rotating baffle steam extraction control performance evaluation method cannot accurately predict the nonlinear behavior of the system response caused by angle changes, and lacks a quantifiable and traceable performance indicator system.

[0016] As a preferred solution of the rotating baffle steam extraction control performance evaluation system described in the present invention, it includes: a data preprocessing and operating condition feature construction module, a nonlinear response modeling and performance index module, and a performance mapping classification and control strategy module; the data preprocessing and operating condition feature construction module is used to perform real-time data acquisition of the rotating baffle steam extraction control operation, and construct an operating condition-state-response data set; the nonlinear response modeling and performance index module is used to construct a nonlinear coupling response prediction model driven by the baffle angle, and output three layers of performance indicators to form a hierarchical performance evaluation structure; the performance mapping classification and control strategy module is used to perform classification and judgment based on performance evaluation, and match the corresponding rotating baffle control parameter set.

[0017] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a method for evaluating the performance of rotary baffle steam extraction control.

[0018] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for evaluating the performance of rotary baffle steam extraction control.

[0019] Beneficial effects of the present invention: The rotary baffle steam extraction control performance evaluation method provided by the present invention realizes the quantitative modeling and performance analysis of the whole process from working condition perception, response prediction to control capability identification by constructing a data-driven rotary baffle steam extraction control performance evaluation method. First, by collecting and preprocessing multi-source real-time operation data, a structured data set covering working condition variables, operating status and system response is constructed, providing high-quality input guarantee for subsequent modeling. Secondly, with the rotating baffle angle as the core driving variable, a nonlinear coupling response prediction model integrating load and temperature disturbance correction factors is constructed, which significantly improves the accurate fitting and response trend prediction capabilities of the control dynamic process. On this basis, a three-level performance indicator system of deviation class, dynamic process class and steady-state stability class is constructed through the deviation between the model output and the measured value, realizing a systematic characterization of steam extraction accuracy, control behavior and system volatility. Furthermore, a control efficiency mapping model is constructed and the characteristic factors are extracted and reduced in dimension in combination with principal component analysis, which effectively improves the evaluation efficiency and model generalization capability. Ultimately, by classifying and identifying performance indicator vectors, the control behavior types are clearly defined and a matching control strategy is output, providing an intelligent decision-making basis for system operation. Overall, this method achieves a high-precision, interpretable, and multi-dimensional evaluation of the rotary diaphragm's control performance, effectively overcoming the technical shortcomings of traditional methods, such as poor dynamic adaptability, coarse evaluation granularity, and difficulty in self-diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 This is an overall flow chart of a method for evaluating the performance of rotary baffle steam extraction control provided by the first embodiment of the present invention.

[0022] Figure 2 This is an overall schematic diagram of a rotary baffle steam extraction control performance evaluation system provided by the third embodiment of the present invention. DETAILED DESCRIPTION

[0023] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0024] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a method for evaluating the performance of a rotary baffle steam extraction control system, comprising:

[0025] S1: Collect real-time data of the rotating diaphragm steam extraction control operation and construct the operating condition-state-response data set through preprocessing.

[0026] Furthermore, data from multiple sources of dynamic parameters closely related to steam extraction control are collected during turbine operation, covering multiple typical load conditions (such as start-up and shutdown, ramping, high-load operation, and sudden changes). Collected parameters include: diaphragm rotation angle, extraction pressure and flow, turbine main steam flow, load, speed, and main steam and extraction temperatures.

[0027] It should be noted that after data collection, time synchronization and missing value repair were performed on the data. Outliers were eliminated using the Z-Score method, denoised using sliding window filtering, and normalized based on the parameter type. This preprocessing resulted in a structured operating condition-state-control response dataset for subsequent modeling.

[0028] It should also be noted that the operating condition-state-control response data set includes operating condition parameters, system state parameters, and control response parameters.

[0029] The operating parameters include load level, operating type identification, steam flow, steam temperature, steam pressure, inlet steam pressure disturbance amplitude and historical load change rate.

[0030] The system status parameters include the current diaphragm angle, the diaphragm angle at the previous moment, the diaphragm angular velocity, the controller state mode, the valve operation frequency, and the steam extraction branch pressure drop.

[0031] The control response parameters include actual extraction steam flow, extraction steam pressure change response, extraction steam response time, maximum overshoot in the response process, extraction steam steady-state mean, response curve fluctuation rate and adjustment period.

[0032] S2: Based on the real-time data of the rotating baffle steam extraction control operation, a nonlinear coupling response prediction model with the rotating baffle angle as the driving variable is constructed. A three-layer performance indicator system is constructed based on the deviation between the nonlinear coupling response prediction model and the actual monitoring value.

[0033] Furthermore, after completing the data collection and preprocessing in step one, based on the collected time series data such as the rotating diaphragm angle θ, extraction pressure ζ, turbine load ψ, and main steam temperature η, a nonlinear coupling model with the rotating diaphragm angle as the main driving variable is constructed to characterize the response characteristics of the extraction steam control system.

[0034] During the modeling process, the actual diaphragm rotation angle θ is used as the primary input variable. Auxiliary operating variables of the current system, including extraction pressure ζ, turbine load ψ, main steam temperature η, and system operating target setpoints (such as target angle α, expected load ρ, and temperature cycle coefficient v), are also introduced as high-order influencing factors. Based on these variables, the predicted extraction response value Q is defined and modeled using the following nonlinear expression:

[0035]

[0036] Where θ is the current angle of the rotating diaphragm, α is the target diaphragm angle setting value, β is the angle adjustment sensitivity factor, ζ is the current extraction pressure, γ is the load disturbance adjustment coefficient, ψ is the current turbine load, ρ is the expected load value, k is the temperature disturbance correction factor, η is the main steam temperature, and v is the steady-state temperature cycle parameter.

[0037] When θ=α, ψ≈ρ, and ζ is high, exp(·)→1, tanh(·)→0, and the overall prediction value Q reaches the peak value Q max , indicating that the extraction response is good; if θ deviates from the target angle, the load disturbance is severe, and the steam temperature change is highly periodic, the model will produce nonlinear attenuation and correction, and Q will decrease.

[0038] The angle deviation term (θ-α) is used to measure the control amplitude difference between the current diaphragm opening and the target state, supplemented by an exponential form to characterize the nonlinear suppression characteristics; ζ is processed by growth compression using a logarithmic function to improve the resolution in the medium and high pressure range; the turbine load offset term (ψ-ρ) is dynamically regularized using a hyperbolic tangent function to prevent nonlinear mutations in the model output under high load disturbances; and the main steam temperature effect is processed using a periodic sinusoidal perturbation function to simulate the thermomechanical coupling changes caused by temperature, especially the indirect impact on the stability of the extraction system when the temperature fluctuates significantly.

[0039] During the modeling process, the parameters in the formula, such as β, γ, κ, etc., are dynamically estimated by fitting historical collected data, and rolling window training is performed on each set of working condition inputs to ensure that the model has good adaptability and generalization capabilities in different load segments and angle ranges.

[0040] S3: A control efficiency mapping model is constructed based on the partition angle and the three-layer performance indicators, and the performance indicator vector output by the control efficiency mapping model is classified and identified.

[0041] Furthermore, based on the nonlinear coupling response prediction model Q, a three-layer substructured performance evaluation index system for the entire steam extraction control process is constructed, corresponding to the three types of response attributes in the control process: execution accuracy, dynamic response characteristics, and steady-state stability.

[0042] This indicator system uses collected observable data such as diaphragm angle θ, extraction pressure ζ, main steam temperature η, load ψ, response time τ, and adjustment frequency δ as its basic variables, and establishes a functional linkage with the output predicted extraction volume ξ = Q(θ, ψ, ζ, η). Ultimately, the following three indicator functions are formed, which are computable, dynamically adaptable, and hierarchically interpretable:

[0043] First layer: The deviation index function Φ1 is used to evaluate the error between the model-predicted steam extraction volume and the measured steam extraction volume. At the same time, the target steam extraction pressure deviation is incorporated into the nonlinear correction, which is expressed as:

[0044]

[0045] The second layer: The dynamic process index function Φ2 is used to describe the response behavior from the start of the diaphragm action to the gradual stabilization of the system. The response intensity caused by the angle change rate, main steam temperature fluctuation and load deviation is expressed as:

[0046]

[0047] The third layer: The steady-state stability index function Φ3 is used to characterize the coupling relationship between the volatility of the steam extraction behavior and the adjustment frequency after the steam extraction system enters the steady state:

[0048]

[0049] Where Φ1 is the execution deviation index value, Φ2 is the dynamic process index value, Φ3 is the steady-state stability index value, ξ is the predicted extraction steam capacity, χ is the measured extraction steam capacity, ∈ is the extraction steam deviation tolerance, ζ is the measured extraction steam pressure, π is the set target extraction steam pressure, ψ is the current load, ρ is the target load value, θ is the baffle angle, η(t) is the curve of the main steam temperature changing with time, v is the temperature cycle adjustment parameter, τ is the dynamic response time period, Q k is the kth steady-state extraction response value, is the steady-state average steam extraction volume, m is the number of steady-state sampling points, and δ is the number of adjustment actions per unit time.

[0050] Φ1∈[0,1] The smaller its value is, the more accurate the model prediction is and the smaller the execution error is; The smaller the value, the weaker the control fluctuation and the smoother the response; The smaller the value, the smaller the steady-state fluctuation and the more stable the system. If δ is high, the exponential term expands, indicating that the system is frequently adjusted and unstable.

[0051] It should be noted that based on the data collected in step 1 and the variables actually used in the model in step 2, the system input factor vector is defined as:

[0052]

[0053] Among them, θ represents the current rotating diaphragm angle, ψ is the current load, ζ is the extraction pressure, η is the main steam temperature, and ρ is the target load setting value, all of which are derived from the system's real-time monitoring data.

[0054] In order to avoid the interference of multicollinearity or dimensional redundancy between factors on the accuracy of analysis, the principal component analysis (PCA) method is used to transform U and extract the characteristic factor combination with the maximum explained variance. The main factor tensor is expressed as:

[0055]

[0056] in, is the orthogonal transformation matrix of the first r principal components selected by eigenvalue decomposition, Enter the high-order nonlinear mapping model as input variables.

[0057] After obtaining the dimension-reduced input Z, a mapping function G(·) is established to calculate and output the performance indicator vector Φ. This mapping combines multiple activation functions, periodic functions, normalization functions, and nonlinear combination relationships, and is expressed as:

[0058]

[0059] in, They are output mapping, cycle amplitude control, logarithmic response enhancement, and frequency adjustment coefficient matrix, which are obtained by historical sample fitting.

[0060] The formula structurally reflects the nonlinear, periodic and coupled mapping rules between input factors and output performance indicators.

[0061] It should also be noted that the grey correlation coefficient matrix is ​​constructed using grey system theory. Output a single indicator dimension γ i Expressed as:

[0062]

[0063] Among them, γ i is the grey correlation degree between the ith comparison sequence and the reference sequence, n is the length of the sequence (i.e. the number of sample moments), k is the kth time point or data index in the sequence, yk is the value of the reference sequence at the kth time point, x ik is the value of the ith comparison sequence at the kth time point, x jk is the value of the jth comparison sequence at the kth time point, ζ is the resolution coefficient used to adjust the relative weight of the maximum difference in the ratio calculation, represents the minimum absolute difference between all sequences and the reference sequence, represents the maximum absolute difference between all sequences and the reference sequence, |y k -x ik | represents the absolute difference between the current comparison sequence and the reference sequence at the kth time point

[0064] By calculating this value for all indicator and factor combinations, the correlation matrix is ​​obtained for sensitivity ranking analysis.

[0065] The output Mapped to the three-dimensional performance space, the control performance is divided into the following four typical states according to the relationship between the high and low index values ​​as shown in Table 1:

[0066] Table 1 Typical status table

[0067]

[0068] Each type is mapped to an optimization strategy vector S = [θ * ,α * ,v * ,r * ] T , which contains four policy outputs, θ * is the target diaphragm angle at the current moment, α * is the maximum angle adjustment range, v * is the dynamic adjustment speed in the response process, r * The minimum refresh period for the steady-state maintenance period. The type-strategy mapping rules are shown in Table 2:

[0069] Table 2 Type-Policy Mapping Rules

[0070]

[0071] Example 2, an embodiment of the present invention, provides a method for evaluating the performance of rotary baffle steam extraction control. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0072] First, in this example, a systematic test of the performance evaluation method for rotary diaphragm extraction control was conducted to address the challenges of multiple load dynamics, inconsistent control accuracy, and unpredictable steady-state fluctuations encountered during the operation of a 600MW steam turbine-generator unit. The test cycle covered the unit's entire process from partial load startup, ramp-up to stable operation, and then sudden regulation to high load. Operational data was collected at several key stages. High-frequency synchronous data acquisition was used. Key parameters recorded included rotary diaphragm angle, extraction pressure, main steam temperature, load value, and extraction response time, covering both system dynamic characteristics and steady-state performance. Samples were collected over multiple complete load cycles. After data cleaning (outlier removal and missing value repair), sliding window filtering was applied. Z-score normalization was performed on various physical quantities based on their dimensions and ranges. Based on this dataset, a nonlinear coupling modeling method driven by diaphragm angle was used to develop response prediction models. A three-tiered performance indicator system was constructed: execution accuracy, response characteristics, and steady-state stability. Based on the distribution of performance indicators in numerical space, a threshold classification method is then used to categorize the system's current control performance into four typical states: stability-first, response-accelerated, precision-compensated, and comprehensive weakening. Each type is associated with a corresponding set of rotary diaphragm control strategy parameters, including target angle, maximum adjustment range, adjustment rate, and refresh period. The strategy output is then combined with measured feedback for correction and update, forming a complete performance-driven strategy recommendation loop.

[0073] Table 3 Experimental data table

[0074]

[0075] The data in the table shows that different samples exhibit distinct performance index structures under the corresponding test conditions. For example, Samples 1 and 2 are categorized as stability-first and response-accelerated, respectively, with extraction rate deviations of 0.12 and 0.18, respectively, placing them in the high-precision category. However, their performance in the response process index and steady-state fluctuation index differs significantly. The former exhibits low response process fluctuation (response process index of 0.22), while the latter exhibits a more intense response process (response process index of 0.67). This suggests that even under high-precision control conditions, system regulation still requires differentiated management. Sample 3, a precision-compensated model, exhibits an extraction rate deviation of 0.35. While its response process and steady-state fluctuations remain within reasonable limits, the resulting deviation suggests that the system requires a small-scale angle compensation and a compression strategy. Sample 4, a comprehensive weakening model, exhibits elevated values ​​in all three indicators, particularly the extraction rate deviation of 0.72 and the response process index of 0.81. This indicates that the system is at the boundary of nonlinear disturbance response under these conditions, making traditional control methods difficult to effectively correct.

[0076] Compared with existing methods based on static threshold control or linear feedback models, the proposed rotating baffle steam extraction control performance evaluation method embodies three innovations: first, by constructing a multidimensional modeling structure for baffle angle + operating condition factors, fine modeling of complex nonlinear response characteristics is achieved; second, by designing a three-layer performance indicator system with interval perception capability, the system control state is made quantifiable, explainable, and classifiable; third, an indicator classification-driven mapping strategy structure is introduced in the strategy generation stage to achieve an upgrade from static control to dynamic matching, effectively avoiding problems such as regulation hysteresis and disturbance amplification under existing methods.

[0077] Overall, the methodology adopted by this invention can truly reflect the subtle differences between operating states and guide the generation and evolution of actual control strategies based on quantitative results, demonstrating structural innovation and engineering adaptability advantages that are significantly different from existing control logic. The automatic generation mechanism of target angles and adjustment amplitudes driven by classification strategies effectively improves system control accuracy and response sensitivity, especially in highly dynamic disturbances or complex working conditions, with stronger stability and recovery capabilities.

[0078] Example 3, reference Figure 2 , which is an embodiment of the present invention, provides a rotating baffle extraction steam control performance evaluation system, including a data preprocessing and operating condition feature construction module 100, a nonlinear response modeling and performance index module 200, and a performance mapping classification and control strategy module 300.

[0079] Among them, S4: data preprocessing and operating condition feature construction module 100 is used to perform synchronous collection of multi-source parameters during the operation of rotating baffle steam extraction control. The collection content includes but is not limited to rotating baffle angle, extraction pressure, extraction flow, main steam temperature, turbine load, speed and other operating condition related variables, and constructs a data set with a three-dimensional structure of operating condition-state-response.

[0080] It should also be noted that the data preprocessing and operating condition feature construction module 100 removes outliers, fills in missing values, and aligns timestamps on the collected data, uses the Z-Score method to filter abnormal deviation samples, and reduces noise through sliding window filtering; then, different types of parameters are classified into three structural domains of "operating condition parameters", "system state parameters" and "control response parameters" according to their physical properties, and after unified coding, a standardized input data structure that can be called for modeling is formed; the preprocessed data set is called by the nonlinear response modeling and performance index module 200 as an input variable group, and as the original data source for subsequent rolling fitting and dynamic index analysis.

[0081] S5: The nonlinear response modeling and performance index module 200 is used to construct a nonlinear coupling response prediction model with the rotating baffle angle as the driving variable. The angle term, pressure term, load term, and temperature disturbance term form a dynamic combination of composite functions such as exponential, logarithmic, tangent, and period, and output the extraction steam prediction value as the response variable. At the same time, based on the error relationship between the model prediction value and the sampled measured value, a three-level performance index system is constructed to evaluate the extraction steam accuracy, response dynamic characteristics, and steady-state volatility respectively.

[0082] It should also be noted that the nonlinear response modeling and performance indicator module 200 performs functional fitting on the core input variables contained in the modeling formula and the system structure characteristics, and adopts a rolling window training strategy to continuously update the model parameters; the three-layer performance indicators constructed are execution deviation indicators, dynamic response indicators and steady-state stability indicators, forming a three-dimensional performance vector structure, which serves as the input condition of the performance mapping classification and control strategy module 300, driving subsequent classification judgment and strategy output.

[0083] S6: The performance mapping classification and control strategy module 300 is used to receive the three-layer performance indicators output by the module 200, perform indicator space mapping and classification identification, and divide the current operating status into categories such as stability priority, response acceleration, accuracy compensation and comprehensive weakening.

[0084] It should also be noted that the performance mapping classification and control strategy module 300 performs spatial clustering or threshold judgment based on the numerical distribution of the performance index vector, and maps the output strategy vector according to the classification result. The strategy vector includes parameters such as the current target partition angle, the maximum angle adjustment amplitude, the adjustment rate, and the refresh cycle. After the strategy is generated, the module 300 caches and records the strategy output, and monitors whether the strategy feedback status meets the performance index improvement conditions. If the feedback continues to fail to meet the standards, a model correction request can be issued to the nonlinear response modeling and performance index module 200, or a re-sampling signal can be issued to the data preprocessing and operating condition feature construction module 100 to trigger the parameter update or data correction process, forming a complete performance-driven control strategy closed-loop mechanism.

Claims

1. A method for evaluating the performance of rotary baffle steam extraction control, characterized in that: include: Collect real-time data on the operation of rotary diaphragm steam extraction control, and construct a working condition-state-response data set through preprocessing; Based on the real-time data of the rotary diaphragm steam extraction control operation, a nonlinear coupling response prediction model with the rotary diaphragm angle as the driving variable was constructed. A three-layer performance indicator system was constructed based on the deviation between the nonlinear coupling response prediction model and the actual monitoring value. A control efficiency mapping model is constructed based on the partition angle and three-layer performance indicators, and the performance indicator vector output by the control efficiency mapping model is classified and identified; The three-tier performance indicator system includes deviation indicators, dynamic process indicators and steady-state stability indicators; The nonlinear coupling response prediction model includes introducing auxiliary operating condition variables to define the predicted extraction response value, building a nonlinear coupling response prediction model, fitting the real-time data of the rotating diaphragm extraction control operation to perform dynamic estimation during the construction process, and predicting the response amount under the driving of the rotating diaphragm angle.

2. The method for evaluating the performance of rotary baffle steam extraction control according to claim 1, wherein: The real-time data of the rotary diaphragm steam extraction control operation is collected, Diaphragm rotation angle, extraction steam pressure and flow, turbine main steam flow, load, speed, main steam temperature and extraction steam temperature.

3. The method for evaluating the performance of rotary baffle steam extraction control according to claim 1 or 2, characterized in that: The pre-processed condition-state-response data set includes: The real-time data of the rotating diaphragm extraction steam control operation are synchronized, missing data are completed, and outliers are eliminated. Moving average filtering and normalization are used to form a structured process condition-state-response data set.

4. The method for evaluating the performance of rotary baffle steam extraction control according to claim 3, wherein: The nonlinear coupling response prediction model based on the real-time data of the rotary diaphragm steam extraction control operation with the rotary diaphragm angle as the driving variable is constructed, including: Based on the acquired rotating diaphragm angle θ, extraction steam pressure ζ, turbine load ψ, and main steam temperature η, a nonlinear coupling model with the rotating diaphragm angle as the main driving variable is constructed, which is expressed as: Where θ represents the current angle of the rotating diaphragm, α represents the target diaphragm angle setting value, β represents the angle adjustment sensitivity factor, ζ represents the current extraction steam pressure, γ represents the load disturbance adjustment coefficient, ψ represents the current turbine load, ρ represents the expected load value, κ represents the temperature disturbance correction factor, η represents the main steam temperature, and v represents the steady-state temperature cycle parameter.

5. The method for evaluating the performance of rotary baffle steam extraction control according to claim 4, characterized in that: The three-layer performance index system constructed by the deviation between the nonlinear coupling response prediction model and the actual monitoring value includes: Based on the deviation between the output value of the nonlinear coupling model and the actual monitoring value, a three-level performance indicator system consisting of execution deviation indicators, dynamic process indicators and steady-state stability indicators is constructed to characterize the steam extraction accuracy, control response behavior and system fluctuation characteristics, respectively.

6. The method for evaluating the performance of rotary baffle steam extraction control according to claim 5, characterized in that: The control efficiency mapping model constructed by using the partition angle and the three-layer performance indicators includes: Determine the input factor vector based on the collected real-time data of the rotating diaphragm extraction steam control operation and the variables actually used in the nonlinear coupling response prediction model; The principal component analysis method is used to transform the input factors, extract the characteristic factor combination with the largest explained variance, output the principal factor tensor, and construct the regulatory effectiveness mapping model to output the mapping relationship.

7. The method for evaluating the performance of rotary baffle steam extraction control according to claim 6, wherein: The classification and identification of the performance indicator vector output by the control effectiveness mapping model includes: According to the mapping relationship output by the control efficiency mapping model, classification and identification are performed, and four performance states are divided into stability priority type, response acceleration type, precision compensation type and comprehensive weakening type. The corresponding rotating diaphragm control strategy parameter vector is output for each performance type.

8. A rotary baffle steam extraction control performance evaluation system, characterized by: It includes a data pre-processing and working condition feature construction module (100), a nonlinear response modeling and performance index module (200), and a performance mapping classification and control strategy module (300); The data preprocessing and operating condition feature construction module (100) is used to perform real-time data collection of the rotary baffle steam extraction control operation and construct an operating condition-state-response data set; The nonlinear response modeling and performance index module (200) is used to construct a nonlinear coupling response prediction model driven by the diaphragm angle, and output three layers of performance indexes to form a hierarchical performance evaluation structure; The performance mapping classification and control strategy module (300) is used to perform classification and discrimination based on performance evaluation, and match corresponding rotary diaphragm control parameter sets.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for evaluating the performance of rotary baffle extraction steam control according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for evaluating the performance of rotary baffle extraction steam control according to any one of claims 1 to 7 are implemented.

Images