Performance evaluation system for turbine speed control system for uncertainty handling

Abstract

This application relates to a performance evaluation system for a turbine speed control system oriented towards uncertainty handling, including an uncertainty source quantification module, a functional evaluation module, a hardware evaluation module, a comprehensive evaluation module, and a result output module. The uncertainty source quantification module is used to identify three types of uncertainty sources in the operation of the turbine speed control system: data noise, operating condition fluctuations, and fault determination ambiguity, and quantify their impact. The functional evaluation module and the hardware evaluation module respectively conduct evaluations with uncertainty correction on the functional execution effect and hardware equipment status of the speed control system. The comprehensive evaluation module integrates the evaluation results of the two modules and dynamically allocates weights. The result output module outputs evaluation conclusions containing credibility and uncertainty source tracing. This application provides scientific support for preventive maintenance and accurate fault location of turbine speed control systems, effectively improves the robustness and engineering adaptability of evaluation results, and ensures the stable operation of the speed control system under complex operating conditions.

Description

Technical Field

[0001] This application relates to the field of performance evaluation technology for turbine speed control systems, and specifically to a performance evaluation system for turbine speed control systems oriented towards uncertainty handling. Background Technology

[0002] Currently, the performance evaluation technologies for speed control systems used in the industry are mostly based on preset static parameters and rigid judgment logic, which are difficult to adapt to the dynamic uncertainties under the complex operating scenarios of hydropower stations and have significant technical limitations: First, the reliability differences in the data acquisition process have not been quantified; second, the impact of dynamic changes in operating conditions on the evaluation threshold has not been adapted; and third, the ambiguity of fault judgment has not been addressed.

[0003] These technological shortcomings lead to frequent "misjudgments" and "missed judgments" in traditional assessment methods: healthy equipment is misjudged as "abnormal" due to data noise, causing unnecessary maintenance and increasing operation and maintenance costs; defective equipment is missed as "normal" due to fluctuations in operating conditions masking risks, which may lead to the expansion of faults, affecting both the stability of power grid supply and reducing the economic efficiency of hydropower station operation. Therefore, developing a performance assessment technology for speed regulation systems that can accurately identify, quantify, and dynamically handle uncertainties has become an urgent need for the industry to cope with complex operating conditions and improve the reliability of assessments. Summary of the Invention

[0004] The purpose of this application is to provide a performance evaluation system for a turbine speed control system oriented towards uncertainty handling, which is specifically designed to solve uncertainty problems such as data noise, operating condition fluctuations, and fuzzy fault determination in the performance evaluation of turbine speed controllers.

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] This application provides a performance evaluation system for a turbine speed control system oriented towards uncertainty handling, including an uncertainty source quantification module, a functional evaluation module, a hardware evaluation module, a comprehensive evaluation module, and a result output module. The uncertainty source quantification module is used to identify three types of uncertainty sources in the operation of the turbine speed control system: data noise, operating condition fluctuations, and fault determination ambiguity, and quantify their impact. The functional evaluation module and the hardware evaluation module respectively conduct evaluations with uncertainty correction on the functional execution effect and hardware equipment status of the speed control system. The comprehensive evaluation module integrates the evaluation results of the two modules and dynamically allocates weights. The result output module outputs evaluation conclusions containing credibility and uncertainty source tracing, thus achieving a robust evaluation of the speed control system performance and solving the deviation problem caused by uncertainty in traditional fixed threshold evaluations.

[0007] The specific implementation of the uncertainty source quantification module includes: quantifying data noise uncertainty through uncertainty coefficient α, where the online monitoring quantity α takes a value of 0.8~1.0, the quasi-online monitoring quantity α takes a value of 0.6~0.8, and the offline monitoring quantity α takes a value of 0.4~0.6, and α is calculated through sensor accuracy, data fluctuation, and cross-validation deviation; quantifying the uncertainty of operating condition fluctuation through a transfer model, establishing the transfer formula Δt = k×ΔH²+b for head fluctuation ΔH and guide vane adjustment time Δt, where k and b are coefficients fitted based on historical data, and establishing a correlation model between load change amplitude and speed overshoot rate; quantifying the fuzzy uncertainty of fault determination through confidence intervals, labeling the fault location result with a confidence level of 0~100%, and triggering secondary verification when the confidence level is below 70%.

[0008] The functional evaluation module's uncertainty-corrected evaluation includes: classifying state quantities into a two-dimensional category of "acquisition type - uncertainty coefficient α"; correcting the state quantity weights based on α, with the correction formula being: corrected weight = original weight × (1 + 0.05 × (α - 0.8)) (α ∈ [0.4, 1.0]); calculating the membership degree of state quantity degradation using triangular fuzzy numbers, such as when the measured value in the moderate degradation range is close to the mild degradation boundary, the deduction value = original deduction value × membership degree; correcting the state quantity score with α, with the formula being: state quantity score = 100 - (basic deduction value × corrected weight) × (1 - α / 2); and taking the weighted average of the state quantity scores for component evaluation. When the lowest score corresponds to α < 0.5, the lowest score weight = α1 / (α1 + α2) and the second lowest score weight = α2 / (α1 + α2), where α1 is the lowest score α and α2 is the second lowest score α.

[0009] The hardware evaluation module's uncertainty-corrected evaluation includes: equipment maintenance cycle correction, introducing a correction coefficient β = failure rate fluctuation δ_f × operating environment uncertainty γ_env, where δ_f is the failure interval variation coefficient, γ_env is the mean value of the operating environment index α, the corrected maintenance cycle = basic maintenance cycle × β, and the expected next maintenance date = current maintenance date + corrected maintenance cycle ± ΔT (ΔT = corrected maintenance cycle × (1 - α_env)); fault state determination adopts multi-condition credibility fusion, calculating the comprehensive credibility of each condition α when entering the maintenance state determination, and determining the entry into the maintenance state when the comprehensive credibility is ≥70%, otherwise triggering low credibility condition review; when determining the status of faulty components, continuous monitoring for 30 minutes after the fault signal disappears and the average α ≥ 0.75 is required to determine that the fault elimination is completed.

[0010] The dynamic weight fusion of the comprehensive evaluation module includes: using α-driven weight allocation, and functional evaluation weights. Hardware evaluation weight , , As the basic weights of AHP, To assess the confidence level for functional evaluation, Assess confidence level for hardware; Overall score = Functional assessment score × + Hardware evaluation score × Introducing a "one-vote veto with flexible deductions," the deduction for minor faults is calculated as 10 × C. fault Major fault deduction = 40 × C fault C fault The confidence level for fault determination.

[0011] The output of the results output module includes: performance scores, overall confidence level, and uncertainty source tracing report for each sub-component of the speed control system; it also supports the output of visual charts, which facilitates maintenance personnel to trace the source of uncertainty and formulate targeted maintenance strategies.

[0012] Compared with existing technologies, the beneficial effects of this invention are: it effectively solves problems such as data noise, operating condition fluctuations, and fuzzy fault judgment in the evaluation of turbine governors; it achieves dynamic fusion by quantifying the reliability coefficient α of multi-source monitoring data; it adopts a dual-module architecture of "function + hardware" and combines a time-series intuitionistic fuzzy model to dynamically correct weights, using α to drive maintenance cycle optimization and reliability fusion fault judgment, breaking through the rigid limitations of traditional methods; it outputs three elements: "score + confidence level + uncertainty source tracing", improving the accuracy and robustness of the evaluation, providing support for preventive maintenance and fault location, and enhancing engineering adaptability and system operation stability. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a block diagram of the performance evaluation system. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. It should be noted that similar reference numerals and letters in the following drawings indicate similar items; therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0016] The terms “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0017] The terms “first,” “second,” etc., are used only to distinguish one entity or operation from another, and should not be construed as indicating or implying relative importance, nor as requiring or implying any such actual relationship or order between these entities or operations.

[0018] This invention relates to an evaluation method based on the "functional evaluation-hardware evaluation-comprehensive evaluation" framework of a water turbine speed control system, which adds "uncertainty characterization, quantification, and correction" logic to each sub-module.

[0019] The evaluation system for the speed control system designed in this scheme mainly consists of four subsystems: the electrical part of the speed controller, the mechanical part of the speed controller, the electrical part of the hydraulic system, and the mechanical part of the hydraulic system. Each subsystem includes functional evaluation and hardware evaluation. For example... Figure 1 As shown:

[0020] Considering the data collection cycle and method for the assessed objects, the status quantities are categorized into three types: online status monitoring quantities, quasi-online status monitoring quantities, and offline status monitoring quantities. Online status monitoring quantities are data from the monitoring system, continuously collected, and with strong real-time characteristics. Quasi-online monitoring quantities are daily inspection data, with slightly weaker real-time characteristics, but can serve as a good supplement to online data. Offline monitoring quantities are monitoring information obtained through testing and maintenance, with weaker real-time characteristics, but a comprehensive evaluation perspective, and can serve as a source of information on the basic status of the equipment. Among these, offline status monitoring quantities provide a more comprehensive reflection of the static health status of the unit during maintenance and shutdown periods, reflecting the basic status of the equipment. Online and quasi-online status monitoring quantities, as real-time inputs, dynamically reflect the health status of the equipment, with offline and online data complementing each other.

[0021] Considering the uncertainty in the data acquisition cycle and method of the state variables of the evaluation object, the state variables are classified into levels, as shown in Table 1:

[0022] Table 1 Classification of Uncertainties in State Quantities

[0023] The established functional evaluation model is divided into two parts: overall evaluation and component evaluation, and all calculations need to incorporate uncertainty correction coefficients.

[0024] The evaluation model categorizes each evaluation object (component) into 10 levels, from least to most important, based on the degree of importance of the state variables to the safe, reliable operation and performance of each evaluation object (component), with corresponding weights of 1 to 10, as shown in Table 2:

[0025] Table 2. Classification of the Importance of State Variables

[0026]

[0027] The uncertainty factor varies depending on the admission status, so an "uncertainty correction term" is added to the original weights. If the uncertainty coefficient α < 0.6 (e.g., offline data), the weight is reduced by 10% (e.g., the original weight was 10, corrected to 9); if α > 0.9 (e.g., online data), the weight is increased by 5% (e.g., the original weight was 8, corrected to 8.4). The correction formula is as follows:

[0028] The corrected weight = original weight × (1 + 0.05 × (α - 0.8)) (α ∈ [0.4, 1.0], to avoid excessive fluctuations in weight).

[0029] Meanwhile, based on the degree of degradation of the state variables, they are divided into 1 to 10 levels from mild to severe, with corresponding deduction values ​​of 1 to 10, as shown in Table 3. The most severe degradation, exceeding the limit of the normal operating range, is scored 7 to 10 points; moderate degradation, within the limit of the normal operating range, is scored 4 to 6 points; and mild degradation, showing a trend towards the limit of the normal operating range but not exceeding the limit, is scored 1 to 3 points.

[0030] Table 3 Classification of Deterioration Degree of State Quantities

[0031]

[0032] The state quantity evaluation has a maximum score of 100 points. The deduction value of a state quantity is determined by both its weight and the basic deduction value. That is, the deduction value of a state quantity equals its basic deduction value multiplied by its adjusted weight. The deduction calculation formula is as follows:

[0033] The score for the state quantity = 100 - basic deduction value × adjusted weight

[0034] The state quantity evaluation score only considers the score of each individual state quantity. The evaluation criteria are categorized as follows:

[0035] 1) When the score of any single state variable reaches the interval [100, 60), it is considered a normal state;

[0036] 2) When the score of any single state variable reaches the interval [60, 40), it is considered an attention state;

[0037] 3) When the score of any single state variable reaches the interval [40, 20), it is considered an abnormal state;

[0038] 4) When the score of any single state variable reaches the interval [20,0], it is considered a severe state.

[0039] The evaluation score of the main component is based on the evaluation score of the state variables, and the lowest evaluation score of the state variables is the evaluation score of the component.

[0040] Table 4 Component Condition Evaluation Table

[0041]

[0042] The overall evaluation score of a subsystem is based on the evaluation scores of its main components, and the lowest evaluation score of a main component is the evaluation score of that subsystem.

[0043] The functional evaluation method focuses on "quantifying uncertainty and accurately reflecting the performance of the speed control system." It is implemented through four steps: "data preprocessing, model building, hierarchical calculation, and result output." The method addresses three types of uncertainty: "data noise, operating condition fluctuations, and judgment ambiguity." The evaluation process is as follows:

[0044] Data preprocessing

[0045] Uncertainty is categorized and quantified to lay a reliable data foundation for the assessment. First, the state variables related to the speed control system function are categorized according to their acquisition method into online monitoring variables (20ms real-time acquisition, α=0.8~1.0), quasi-online monitoring variables (daily inspection, α=0.6~0.8), and offline monitoring variables (quarterly testing, α=0.4~0.6). The reliability of each type of data is quantified using the uncertainty coefficient α. Then, data uncertainty is addressed specifically: online data is filtered for noise jumps using the 3σ criterion; quasi-online data is treated with a time-series decay factor to mitigate lag effects; and offline data serves as a static baseline, cross-validated with online data to ensure that the basic data input for assessment has been initially filtered for uncertainty interference.

[0046] Model building

[0047] By embedding uncertainty correction logic, the adaptability bias of traditional fixed models is avoided. On the one hand, a two-level index system of "component-subsystem" is constructed, with the indexes taking into account both functional core and uncertainty adaptation. At the component level, the focus is on key functional units, and the indexes include dynamic thresholds; at the subsystem level, component indexes are integrated to cover the overall function of the speed control system, and each index is associated with the α value of the corresponding state variable, clarifying the source of data reliability. On the other hand, a "combined weight model" is designed to overcome the limitations of single AHP subjective weights: first, expert subjective weights are calculated using AHP, then objective weights are calculated based on data dispersion using the entropy weight method (the more dispersed the data, the lower the α, and the smaller the entropy weight), and finally, the α mean is introduced to correct the combined weights, ensuring that the weight allocation reflects expert experience and can be dynamically adjusted due to data uncertainty.

[0048] Hierarchical calculation

[0049] The three-tiered scoring system progressively corrects the results, mitigating the impact of uncertainty on the outcome layer by layer. The first step is to calculate the state quantity score. Based on "100 - basic deduction value × weight", the score is corrected using α (formula: state quantity score = 100 - (basic deduction value × corrected weight) × (1 - α / 2)). The more reliable the data (the higher the α), the closer the deduction is to the actual degree of degradation. The second step is to calculate the component score. Based on the state quantity score, if the α corresponding to the lowest score is <0.5 (the data is extremely unreliable), then "directly taking the lowest score" is abandoned, and instead a weighted average of "lowest score × α1 / (α1+α2) + second lowest score × α2 / (α1+α2)" is used (α1 is the lowest score α, α2 is the second lowest score α), balancing "warning" and "reliability". The third step is to calculate the subsystem score, taking the weighted average of the scores of its components (the weight is the component's functional importance × the mean of the component's score α). If a component's score α <0.6, its weight is appropriately reduced to avoid a single unreliable component score dominating the subsystem evaluation.

[0050] Output

[0051] The output includes subsystem function scores (0~100 points), function assessment confidence (calculated by weighting each state variable α, e.g., confidence = Σ(α × weight) × 100%), and uncertainty tracing report. At the same time, the function assessment results are linked with the hardware assessment. If the function assessment confidence is <70%, the subsequent comprehensive assessment will prompt "the reliability of hardware sensor data needs to be verified first", ensuring that the function assessment can not only independently reflect the functional status of the speed control system, but also provide support for the uncertainty tracing of the overall assessment.

[0052] The hardware evaluation methodology revolves around the hardware equipment of the turbine speed control system, with "full lifecycle data management + dynamic uncertainty correction" as its core, and achieves closed-loop evaluation through four core modules. The evaluation process is as follows:

[0053] Basic data standardization management

[0054] First, a unique coding rule is established through the "Equipment Structure Module" to ensure that the equipment is unique and traceable. Then, the "Basic Information Module" records the equipment model, rated life, commissioning date, and usage environment, links multi-source monitoring data, and ensures the reliability of the basic data for assessment through data consistency verification (marked as "to be reviewed" when the matching degree is lower than α=0.9).

[0055] Dynamic adjustment of maintenance cycle

[0056] Based on the "Equipment Maintenance Date Calculation / Management Module," the fixed cycle model is broken: First, a correction coefficient β (β = failure rate fluctuation δ_f × operating environment uncertainty γ_env, where δ_f is the coefficient of variation of the failure interval and γ_env is the mean value of the environmental indicator α) is introduced to correct the basic maintenance cycle to "corrected cycle = basic cycle × β"; then, the "expected next maintenance date" is dynamically calculated (expected next maintenance date = current maintenance date + corrected cycle ± ΔT, where ΔT is the uncertainty buffer period, determined by the environmental α value); after the fault is repaired, if the repair effect verifies α ≥ 0.8, the cycle is recalculated normally; if α < 0.8, the cycle is shortened by 10% and monitoring is strengthened to avoid inaccurate maintenance due to operating condition fluctuations and data deviations.

[0057] Accurate fault status assessment

[0058] Based on the "Equipment Fault Management Module": First, the fault information of the fault diagnosis module is collected, and the "fault-equipment" correspondence is generated by associating it with the equipment ledger, recording the fault date, cause, level and judgment confidence level; then the status of the faulty component is determined. "Requires elimination" requires that the fault signal lasts for ≥5 sampling cycles and the associated data α≥0.7. "Elimination completed" requires that the fault signal disappears after the unit is turned on, the monitoring data α≥0.75 for 30 consecutive minutes and the fluctuation ≤5%, so as to avoid misjudgment due to signal noise and ambiguous positioning.

[0059] Hardware status quantitative assessment

[0060] Quantification is achieved through the "Equipment Status Assessment Module": Seven assessment indicator dimensions are constructed: "Pre-commissioning status, quality level, family quality history, failure rate, maintenance records, operating time, and operating environment". Each indicator is corrected for deductions using an α coefficient. The weights are combined using "AHP - Entropy Weight - α Correction" to avoid bias from a single subjective / objective weight. The hardware score and confidence level are calculated through an algorithm. The system is classified into "Normal (85~100 points), Attention (70~84 points), Abnormal (60~69 points), and Severe (0~59 points)" levels, with each level marked with an uncertainty range (e.g., Attention status ±8~15 points, determined by the confidence level). The final output is "Score + Confidence Level + Uncertainty Source Tracing".

[0061] The comprehensive evaluation focuses on "integrating functional and hardware evaluation results and quantifying uncertainties throughout the process." It is achieved through five steps: "basic preparation, dynamic weight allocation, score fusion, flexible correction, and result output." This ensures that the evaluation results reflect both the overall performance of the speed control system and clearly identify the impact of uncertainties. The specific methods are as follows:

[0062] Assessment of basic preparations

[0063] First, extract the core outputs of the functional evaluation and hardware evaluation.

[0064] Functional assessment: Subsystem functional score, functional assessment confidence level (Calculated by weighted average of the state variables α in the functional assessment)

[0065] Hardware evaluation: Subsystem hardware score, hardware evaluation confidence level (Calculated by weighted average of various indicators α in the hardware evaluation)

[0066] At the same time, uncertainty source information for both types of assessments will be collected as a basis for subsequent weight adjustments and result source tracing.

[0067] Dynamic weight allocation

[0068] Breaking away from the traditional fixed-weight model, the weight allocation is driven by the "credibility of evaluation results," and the steps are as follows:

[0069] Step 1: Use the Analytic Hierarchy Process (AHP) to determine the basic weights for functional assessment. ) and the basic weights of hardware evaluation ( )

[0070] Step 2: Introduce credibility correction, the calculation formula is: ,

[0071] The higher the credibility of the evaluation results, the higher their weight should be, so as to avoid low credibility results dominating the overall evaluation.

[0072] Core scoring fusion

[0073] Based on dynamic weights, a weighted average of the functional score and the hardware score is used to obtain the initial overall score:

[0074] Initial overall score = Functional score Hardware score

[0075] Elastic correction

[0076] To address real-time fault scenarios and avoid distorted results due to rigid scoring, an "elastic scoring based on fault confidence" approach is adopted:

[0077] First, determine the fault level (minor fault, major fault), and then determine the confidence level C for the fault determination. fault (Minor fault C) fault =65%, Major Fault C fault = 90%

[0078] Then calculate the flexible deduction value:

[0079] Minor fault penalty points = 10 × C fault

[0080] Major fault penalty points = 40 × C fault

[0081] Final overall score = Initial overall score – Flexible deduction value.

[0082] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A performance evaluation system for a hydro-turbine governing system oriented to uncertainty processing, characterized by, The system includes an uncertainty source quantification module, a functional evaluation module, a hardware evaluation module, a comprehensive evaluation module, and a result output module. The uncertainty source quantification module identifies three types of uncertainty sources in the operation of the turbine speed control system: data noise, operating condition fluctuations, and fault determination ambiguity, and quantifies their impact. The functional evaluation module and the hardware evaluation module conduct evaluations with uncertainty correction on the functional performance of the speed control system and the status of the hardware equipment, respectively. The comprehensive evaluation module integrates the evaluation results of the two modules and dynamically allocates weights. The result output module outputs evaluation conclusions containing credibility and uncertainty source tracing. Overall, the system achieves a robust evaluation of the speed control system performance and solves the deviation problem caused by uncertainty in traditional fixed threshold evaluations.

2. The performance evaluation system for the water turbine governing system oriented to the uncertainty processing according to claim 1, characterized in that, The specific implementation of the uncertainty source quantification module includes: quantifying data noise uncertainty through uncertainty coefficient α, where the online monitoring quantity α takes a value of 0.8~1.0, the quasi-online monitoring quantity α takes a value of 0.6~0.8, and the offline monitoring quantity α takes a value of 0.4~0.6, and α is calculated through sensor accuracy, data fluctuation, and cross-validation deviation; quantifying the uncertainty of operating condition fluctuation through a transfer model, establishing the transfer formula Δt = k×ΔH²+b for head fluctuation ΔH and guide vane adjustment time Δt, where k and b are coefficients fitted based on historical data, and establishing a correlation model between load change amplitude and speed overshoot rate; quantifying the fuzzy uncertainty of fault determination through confidence intervals, labeling the fault location result with a confidence level of 0~100%, and triggering secondary verification when the confidence level is below 70%.

3. The performance evaluation system for the water turbine governing system oriented to the uncertainty processing according to claim 1, characterized in that, The functional evaluation module's uncertainty-corrected evaluation includes: classifying state quantities into a two-dimensional category of "acquisition type - uncertainty coefficient α"; correcting the state quantity weights based on α, with the correction formula being: corrected weight = original weight × (1 + 0.05 × (α - 0.8)) (α ∈ [0.4, 1.0]); calculating the membership degree of state quantity degradation using triangular fuzzy numbers, such as when the measured value in the moderate degradation range is close to the mild degradation boundary, the deduction value = original deduction value × membership degree; correcting the state quantity score with α, with the formula being: state quantity score = 100 - (basic deduction value × corrected weight) × (1 - α / 2); and taking the weighted average of the state quantity scores for the component evaluation. When the lowest score corresponds to α < 0.5, the lowest score weight = α1 / (α1 + α2) and the second lowest score weight = α2 / (α1 + α2), where α1 is the lowest score α and α2 is the second lowest score α.

4. The performance evaluation system for a turbine speed regulation system oriented towards uncertainty handling according to claim 1, characterized in that, The hardware evaluation module's uncertainty-corrected evaluation includes: equipment maintenance cycle correction, introducing a correction coefficient β = failure rate fluctuation δ_f × operating environment uncertainty γ_env, where δ_f is the failure interval variation coefficient, γ_env is the mean value of the operating environment index α, the corrected maintenance cycle = basic maintenance cycle × β, and the expected next maintenance date = current maintenance date + corrected maintenance cycle ± ΔT (ΔT = corrected maintenance cycle × (1 - α_env)); fault state determination adopts multi-condition credibility fusion, calculating the comprehensive credibility of each condition α when entering the maintenance state determination, and determining the entry into the maintenance state when the comprehensive credibility is ≥70%, otherwise triggering low credibility condition review; when determining the status of faulty components, continuous monitoring for 30 minutes after the fault signal disappears and the average α ≥ 0.75 is required to determine that the fault elimination is completed.

5. The performance evaluation system for a turbine speed regulation system oriented towards uncertainty handling according to claim 1, characterized in that, The dynamic weight fusion of the comprehensive evaluation module includes: using α-driven weight allocation, and functional evaluation weights. Hardware evaluation weight , , As the basic weights of AHP, To assess the confidence level for functional evaluation, Assess confidence level for hardware; Overall score = Functional assessment score × + Hardware evaluation score × Introducing a "one-vote veto with flexible deductions," the deduction for minor faults is calculated as 10 × C. fault Major fault deduction = 40 × C fault C fault The confidence level for fault determination.

6. The performance evaluation system for a turbine speed regulation system oriented towards uncertainty handling according to claim 1, characterized in that, The output of the results output module includes: performance scores, overall confidence level, and uncertainty source tracing report for each sub-component of the speed control system; it also supports the output of visual charts, which facilitates maintenance personnel to trace the source of uncertainty and formulate targeted maintenance strategies.

Images