Reliability evaluation method and device for key component of water turbine governor and medium

Abstract

The invention belongs to the technical field of reliability evaluation of power system equipment, and provides a reliability evaluation method and equipment for key components of a water turbine governor and a medium, and the method comprises the steps: screening out the key components of the water turbine governor through employing an analytic hierarchy process; processing an original fault operation and maintenance record of the key component, and obtaining a data set containing censored data; calculating a deletion rate based on the data set; and dynamically selecting an optimal method according to different censoring rates to complete reliability evaluation of the key component. According to the method, self-adaptive selection of the parameter estimation method is realized through censoring rate interval division, the adaptability to data with different censoring degrees is improved, and the precision and robustness under different censoring rates can be considered.

Description

Technical Field

[0001] The present invention belongs to the technical field of reliability assessment of power system equipment. Specifically, it relates to a reliability assessment method, device and medium for key components of a hydraulic turbine governor, and is applicable to life modeling, reliability analysis and maintenance decision optimization of key components such as complex mechatronic systems and hydraulic turbine unit equipment. Background Art

[0002] As the core control equipment of a hydropower system, the hydraulic turbine governor realizes unit speed control, power regulation and grid frequency stability by adjusting the guide vane opening and blade angle in real time. Its key components include precision mechatronic components such as an electro-hydraulic converter, a PID (Proportional Integral Derivative) controller, and a servomotor, which are long-term subjected to the coupling action of multiple stresses such as hydraulic pulsation, mechanical wear and electrical aging. Once a key component of the governor fails, it will directly lead to increased unit oscillation, decreased grid connection stability and even unplanned shutdown, causing significant economic losses. At present, the records of failure data mainly include the failure occurrence time, failure location, failure cause, failure severity, failure handling situation, etc. Due to the long operation life and high reliability of the governor, there are some components that have not experienced functional failures from the time of commissioning to the end of the observation period statistics. The recording of failure data has the characteristics of a right-censored test at a fixed time, and there is generally a high censoring feature.

[0003] The so-called reliability of a product refers to the ability of the product to complete the specified function under the specified conditions and within the specified time. The analysis of reliability includes quantitative and qualitative analysis. Qualitative analysis mainly analyzes the change of system reliability from the perspectives of failure mechanism, failure mode, failure consequence, etc., and is generally related to the specific research object. Quantitative analysis mainly quantitatively discusses the reliability of the system from quantitative indicators such as statistical analysis of failure data, estimation of reliability, and failure rate function. The reliability analysis process can provide necessary guidance for reliability modeling, including aspects such as modeling methods and model selection. For the research of actual cases, establishing an accurate reliability model is the premise for formulating a reasonable maintenance strategy. This part will apply reliability theory for quantitative analysis, study and determine the theoretical optimal regular maintenance cycles of various components of the unit to meet the personalized needs of users, so as to assist in making decisions on the timing of maintenance.

[0004] In the reliability assessment of complex systems, the Weibull distribution is widely used in life modeling due to its flexibility. Since there are significant differences in the censoring levels of the datasets obtained for different key components of a hydro turbine governor during operation, resulting from different operating characteristics, service life, and reliability levels, it is difficult to assess the reliability of all key components using a single evaluation method. Therefore, a complete reliability assessment solution is needed that can dynamically select the optimal parameter method according to the censoring rate level. Moreover, field data often exhibits the characteristics of "large samples and high censoring", and traditional parameter estimation methods have the following problems: In the case of high-censoring data, the Maximum Likelihood Method (MLM) has a significant estimation bias for the shape parameter ( β ), and the Mean Time to Failure (MTTF) is overestimated; the Least Squares Method (LSM) relies on the fitting of the empirical distribution function, is sensitive to the censoring time distribution, and has large error fluctuations; the Bayesian method requires prior information, the heuristic algorithm has high complexity, and none of them systematically quantifies the impact of censoring characteristics. Therefore, there is an urgent need for a reliability assessment method that can adopt a hierarchical and multi-strategy approach, comprehensively consider censoring characteristics, and effectively utilize censoring information to improve the estimation accuracy and robustness under high-censoring data. Through patent retrieval of existing technologies, the common methods are mainly as follows: Method 1: Chinese Patent Publication No.: CN112733281A, Patent Name: A Machine Tool Reliability Assessment Method Considering Censored Truncated Data. This patent discloses a machine tool reliability assessment method considering censored truncated data to improve the accuracy of reliability assessment parameters. First, normal data is screened based on the IQR (Interquartile Range) method for the data. Secondly, the truncated data in the data is reduced using the total failure time processing rule to expand the sample size of failure data. Then, the initial values of the Weibull parameters are initially obtained using the average rank method. Then, based on the new empirical distribution function, the least squares method is used to complete the estimation of the new Weibull parameters. Finally, the K-S method is used for convergence testing. Method 1 focuses on the processing of truncated data and sample size expansion, and improves the accuracy of machine tool reliability assessment through multi-stage statistical methods.

[0005] Method 2: Chinese Patent Publication No.: CN112507487A, Patent Name: A Reliability Evaluation Method and System for the Servomotor of a Hydraulic Turbine Governor. This patent provides a reliability evaluation method and system for the servomotor of a hydraulic turbine governor, including: learning the mapping relationship between the state index of the servomotor of the hydraulic turbine governor and the action rate of the servomotor based on a neural network and historical data of the hydraulic turbine; determining the action rate of the servomotor at any moment based on the state index of the servomotor of the hydraulic turbine governor at any moment and the mapping relationship; determining the function value of the servomotor based on the action rate of the servomotor at any moment and a preset minimum rate limit value of the servomotor; determining the reliability of the servomotor based on the distribution of the function value of the servomotor in the standard normal coordinate system; the reliability reflects whether the operating state of the servomotor meets the requirements. The present invention effectively fills the gap in the field of reliability evaluation of the servomotor of a hydraulic turbine governor, can obtain a more accurate action rate value, and makes the calculated reliability closer to the actual system reliability. Method 2 focuses on dynamic mapping modeling based on a neural network and function value analysis, filling the gap in the field of reliability evaluation of the hydraulic turbine servomotor.

[0006] Method 3: Chinese Patent Publication No.: CN111291486A, Patent Name: A Reliability Evaluation Method for Components of a Numerical Control Machine Tool System. This patent relates to a reliability evaluation method for components of a numerical control machine tool, including the following steps: 1. Divide the system components, collect on-site fault information of the numerical control machine tool and conduct fault analysis; 2. Calculate the equivalent fault interval time and equivalent test censoring time of the components and the machine tool system; 3. Apply Johnson's method to correct the rank of the equivalent fault interval time and construct the reliability models of the components and the machine tool system; 4. Construct a system series reliability model and apply the correlation index method to verify the rationality of the component reliability modeling based on the equivalent sample method; Under the assumption of "as good as new" repair, the present invention calculates the component fault interval time using the equivalent sample method, which conforms to the definition of the component life of the system. The total fault time method and the equivalent sample method are comprehensively used to correct the component fault interval time, expanding the sample size, which conforms to the sampling principle. Compared with the traditional method of component reliability modeling based on system information, it is more in line with engineering practice. Method 3 focuses on component-level reliability modeling and system-level verification, expands the sample size through the equivalent sample method, and improves the engineering applicability of the evaluation. Summary of the Invention

[0007] The present invention provides a reliability evaluation method, device and medium for key components of a hydraulic turbine governor, and proposes a phased parameter estimation strategy for high-censored data scenarios, aiming to solve the problems of large parameter estimation deviation and poor robustness of traditional methods under high-censored data.

[0008] In a first aspect, the present invention provides a reliability evaluation method for key components of a hydraulic turbine governor, including: Use the analytic hierarchy process for the hydroturbine governor to screen out key components; Process the original fault operation and maintenance records of the key components to obtain a dataset containing censored data; Calculate the censoring rate based on the dataset; Dynamically select the optimal method according to different censoring rates to complete the reliability assessment of key components.

[0009] In some embodiments, the use of the analytic hierarchy process for the hydroturbine governor to screen out key components includes: Establish a hierarchical structure model according to the system structure composition of the hydroturbine governor, and construct a judgment matrix according to the hierarchical structure model; Use the consistency index to test the consistency of the judgment matrix; After meeting the consistency, screen out key components by calculating the component weights.

[0010] In some embodiments, the processing of the original fault operation and maintenance records of the key components to obtain a dataset containing censored data includes: In the provided original fault operation and maintenance records, screen out the original fault operation and maintenance records of the key components according to the component name and process them; In the original fault operation and maintenance records of the key components, after normalizing the component commissioning date and up to the end of the observation time, the obtained component failure data conforms to a random censoring test, thereby obtaining the number of failed components and the component failure time. The running time of the components that did not fail is right-censored, thereby obtaining the censoring time; since the life data of the governor components follows a Weibull distribution, after removing the outliers in the failure time and censoring time, finally obtain a dataset containing censored data.

[0011] In some embodiments, the calculation of the censoring rate based on the dataset includes:

[0012] Where, c is the censoring rate, n is the number of failed components, m is the number of components that did not fail.

[0013] In some embodiments, the dynamic selection of the optimal method according to different censoring rate levels to complete the reliability assessment of key components includes: For low censoring rates, use the maximum likelihood estimation method to complete the reliability assessment of key components; For high censoring rates, use the step-by-step parameter estimation method that combines censoring characteristic information to complete the reliability assessment of key components; For an extremely high censoring rate, the reliability assessment of key components is completed by combining a semi-supervised data conversion method and a maximum likelihood estimation method; Among them, the low censoring rate, high censoring rate, and extremely high censoring rate are determined according to a preset censoring rate range.

[0014] In some embodiments, the reliability assessment of key components completed by using the maximum likelihood estimation method includes:

[0015] Wherein, L is the likelihood function, n is the number of failed components, m is the number of components that have not failed, t i represents the i failure time of the th failure sample, j is the censoring time of the F ( t ) is the shape parameter of the Weibull distribution β and the scale parameter η failure distribution function, expressed as:

[0016] f ( t ) is the shape parameter of the Weibull distribution β and the scale parameter η density distribution function, expressed as:

[0017] Wherein, , t is a random variable, that is, the failure time or censoring time, and exp() is the exponential function with the natural constant e as the base.

[0018] In some embodiments, the reliability assessment of key components completed by using a step-by-step parameter estimation method that combines censoring feature information includes: Calculate censoring feature statistics, including the coefficient of variation of censoring time and the skewness of censoring time; the coefficient of variation of censoring time is the ratio of the standard deviation of censoring time to the mean; it is obtained based on the censoring time and the standard deviation and mean of censoring time; Construct a correction function based on the censoring rate and the skewness of censoring time, and use the correction function to correct the mean time to failure MTTF; wherein, the correction function is a function based on the censoring rate and the skewness of censoring time; For the corrected mean time to failure MTTF, perform iterative optimization based on the maximum likelihood estimation method: Step 1: Fix the estimated value of the mean time to failure (MTTF) corrected by the correction function, substitute it into the calculation formula of the MTTF, and substitute the calculated scale parameter into the method of maximum likelihood estimation to estimate the shape parameter; Step 2: Fix the estimated shape parameter, use the method of maximum likelihood estimation to estimate the scale parameter, and update the estimated value of the MTTF using the estimated scale parameter; Step 3: Fix the updated estimated value of the MTTF, re-estimate the shape parameter according to Step 1, and determine whether the re-estimated shape parameter meets the convergence condition. If it does not meet the condition, repeat Steps 1 to 2 until the convergence condition is met.

[0019] In some embodiments, the reliability assessment of key components is completed by combining the semi-supervised data conversion method and the maximum likelihood estimation method, including: Introduce the operating condition parameters of the turbine governor components into the dataset and use them as the feature vectors of the dataset samples; the dataset includes failure samples and censored samples; Construct a weight matrix by combining feature similarity and survival curve similarity; Calculate the transition matrix based on the weight matrix; Construct an initial label matrix based on the failure time of the failure samples and the censoring time of the censored samples; Iteratively update the pseudo-survival time based on the transition matrix and the initial label matrix to obtain the optimal pseudo-labels of the censored samples; Replace the censored samples in the dataset with the optimal pseudo-labels and regard them as failure samples, thereby updating the dataset; Based on the updated dataset, complete the reliability assessment of key components using the method of maximum likelihood estimation.

[0020] In a second aspect, the present invention provides an electronic device, including: At least one processor; and a memory communicatively connected to the at least one processor; Wherein, the memory stores instructions executable by the at least one processor, and the at least one processor, by executing the instructions stored in the memory, causes the at least one processor to execute the above method.

[0021] In a third aspect, the present invention provides a computer-readable storage medium, which is used to store instructions that, when executed, implement the above method.

[0022] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are: 1. Since there are significant differences in the censoring levels of the obtained datasets for different key components of the hydraulic turbine governor during operation due to their different operating characteristics, service life, and reliability levels, it is difficult to evaluate the reliability of all key components using a single evaluation method. In response to this problem, the present invention can dynamically select the optimal method according to the censoring rate level. For example, when the censoring rate < 40%, the maximum likelihood estimation method (MLE) is used; when the censoring rate is 40% - 80%, a stepwise parameter estimation method that combines censoring characteristic information is used; when the censoring rate > 80%, a semi-supervised data transformation method is combined with the maximum likelihood estimation method. The present invention realizes the adaptive selection of parameter estimation methods through censoring rate interval division, improves the adaptability to data with different censoring degrees, and can also take into account the accuracy and robustness under different censoring rates. Through the above hierarchical and multi-strategy solution, the statistical rigor and method feasibility can be effectively balanced in engineering practice, ensuring the accuracy of the reliability assessment of the hydraulic turbine governor.

[0023] 2. The stepwise parameter estimation method that combines censoring characteristic information proposed in the present invention is applicable to high-censoring scenarios, such as when the censoring rate is 40% - 80%. This method can comprehensively reflect the censoring characteristics of the data by combining multiple statistical features (such as the censoring rate, coefficient of variation, and skewness of the censoring time), and use the statistical information of the censored data to correct the initial estimated value, making full and effective use of the implicit characteristic information in the censored data. Through staged optimization, the coupling problem caused by simultaneously estimating multiple parameters is avoided. This method effectively solves the bias problem of traditional methods under high censoring rates, significantly improving the estimation accuracy and robustness under high-censoring data.

[0024] 3. For extremely high-censoring data, such as when the censoring rate > 80%, traditional parameter estimation methods fail. The present invention introduces a new method of combining survival curve difference measurement and dynamic weight combination in the framework of semi-supervised learning, combines feature similarity and survival time distribution similarity through weight coefficients when constructing the weight matrix, and proposes a semi-supervised data transformation method for survival analysis. This method generates multiple pseudo-candidate times, i.e., pseudo-labels, for the survival time of censored samples through iterative search, and then optimally selects the candidate pseudo-label that minimizes the objective loss function as the optimal pseudo-label for the censored samples, making it as close as possible to the true failure time. Gradually transform the censored samples into failure samples, and finally use the maximum likelihood estimation method to estimate the parameters for the dataset after merging the transformed data and the original failure data. This method can adjust the similarity weight ratio of features and survival time distribution according to the data distribution to adapt to different high-censoring scenarios, and the model has strong adaptability. The method of combining semi-supervised data transformation with maximum likelihood estimation significantly overcomes the limitations of traditional methods in high-censoring data. Its core advantages are: (1) Driven by data, without the need for strong model assumptions, it can flexibly capture the characteristics of the survival distribution.

[0025] (2) Maximize information utilization: Integrate the features of censored data and survival time information, effectively utilize the information of failure samples to infer the survival time of censored samples, thereby improving the prediction performance of the model under high-censored data.

[0026] (2) Robustness: Reasonably define the similarity metric and iterative update rules to ensure that the generated pseudo-labels not only conform to the data distribution but also satisfy the physical constraints of the survival time. It is particularly suitable for scenarios where failure data is scarce but censored data is abundant (such as high-censored data of hydroturbine governors). By intelligently utilizing the potential information of censored data, it significantly improves the model performance and is an ideal choice for fields such as industrial predictive maintenance and medical diagnosis, providing a new solution for modeling high-censored data in fields such as industrial reliability analysis. Description of the Drawings

[0027] Figure 1 It is a flowchart of a reliability evaluation method for key components of a hydroturbine governor provided by an embodiment of the present invention.

[0028] Figure 2 It is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed Embodiments

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated herein can be arranged and designed in various different configurations.

[0030] Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] Embodiment To solve the problems of large parameter estimation deviation and poor robustness of traditional methods under high-censored data. As Figure 1 shown, an embodiment of the present invention provides a reliability evaluation method for key components of a hydroturbine governor, including the following steps: S100, Use the analytic hierarchy process to screen out key components of the hydroturbine governor.

[0032] In some embodiments, step S100 includes the following sub-steps: S101. Build a hierarchical structure model according to the system structure composition of the hydraulic turbine governor, and construct a judgment matrix based on the hierarchical structure model. The size of the judgment matrix is n × n , where n is the number of devices at the corresponding level.

[0033] S102. Use the consistency index to test the consistency of the judgment matrix. In this embodiment, the consistency ratio CR is used as the consistency index. When the consistency ratio CR < CR TH ( CR TH is the preset threshold of the consistency ratio, which is set according to needs and actual situations. For example, when it is 0.1, it is considered that the consistency of the judgment matrix is acceptable; when the consistency ratio CR ≥ CR TH , the judgment matrix is appropriately corrected.

[0034] S103. After meeting the consistency, screen the key components by calculating the component weights. Specifically, calculate the weight vector w corresponding to each level of components, so as to obtain the weight corresponding to each component. The higher the component weight, the more likely the component is to fail relatively, and the more worthy of attention when making maintenance decisions. The components with higher weights (relatively higher weights or exceeding the preset weight threshold w TH ) are regarded as key components for the next component reliability assessment.

[0035] S200. Process the original fault operation and maintenance records of the key components to obtain a data set containing censored data.

[0036] Taking the hydraulic turbine governor as the analysis object, mainly study the failure conditions of the key components under the hydraulic turbine governor. In the provided original fault operation and maintenance records, screen out the original fault operation and maintenance records of the key components according to the component name and process them.

[0037] In the original fault operation and maintenance records of the key components, since the component commissioning dates are different, after normalizing the component commissioning dates to t = 0, until the end of the observation time, the obtained component failure data conforms to the random censoring test, and obtain the number of failed components n and the component failure time t i (the time when the component fails during actual operation), and those that have not failed mThe operating time of the unit components is right-censored, and the censoring time (i.e., the right-censoring time, which will be uniformly described as the censoring time for convenience in the following) is denoted as ; Since the life data of the governor components follow the Weibull distribution, after removing the outliers in the failure times and censoring times, a dataset containing censored data is finally obtained, which is expressed as: (1) where t i represents the component failure time of the i th failed component (i.e., the failure sample in the dataset), represents the censoring time of the operating time of the j th component that has not failed (i.e., the censored sample in the dataset).

[0038] S300, calculate the censoring rate based on the dataset.

[0039] In this embodiment, based on the number of failed components, the number of components that have not failed, and the total number of key components, the formula for calculating the censoring rate c is as follows: (2) where n is the number of failed components, m is the number of components that have not failed, n + m = N , N is the total number of key components.

[0040] The censoring rate level can be determined according to the preset censoring rate range. Among them, the preset censoring rate range is set according to needs and actual situations. The settings in this embodiment are as follows: Low censoring rate: censoring rate < 40%; High censoring rate: censoring rate is in the range of 40% - 80%; Extremely high censoring rate: censoring rate > 80%.

[0041] S400, dynamically select the optimal method according to different censoring rate levels to complete the reliability assessment of key components.

[0042] S410, for the low censoring rate with a censoring rate < 40%, use the maximum likelihood estimation method to complete the reliability assessment of key components. Specifically as follows: The maximum likelihood estimates of the shape parameter β (shape parameter) and scale parameter η (scale parameter) of the Weibull distribution are obtained by maximizing the log-likelihood function of Equation (5).

[0043] Failure distribution function of two-parameter Weibull distribution F ( t ) is as follows: (3) The probability density function (PDF) of the two-parameter Weibull distribution is as follows: (4) where , is the scale parameter (or characteristic life) of the Weibull distribution, and exp() is the exponential function with the natural constant e as the base; is the shape parameter of the Weibull distribution; t is a random variable, i.e., the failure time or censoring time.

[0044] (5) S420. For the high censoring rate with the censoring rate ranging from 40% to 80%, a step-by-step parameter estimation method that combines censoring characteristic information is used to complete the reliability assessment of key components.

[0045] By combining multiple statistical characteristics (such as censoring rate, coefficient of variation of censoring time, and skewness of censoring time), the censoring characteristics of the data are comprehensively reflected, and the initial estimated value is corrected using the statistical information of the censored data. Through staged optimization, the coupling problem caused by simultaneously estimating multiple parameters is avoided. The specific steps of the step-by-step parameter estimation method are as follows: S421. Calculate the censoring characteristic statistics, including the coefficient of variation of censoring time and the skewness of censoring time.

[0046] (1) Calculate the coefficient of variation of censoring time ρ : The coefficient of variation of censoring time is the ratio of the standard deviation of censoring time to the mean, reflecting the dispersion degree of censoring time. Its calculation formula is as follows: (6) where σ c is the standard deviation of censoring time, reflecting the measure of the dispersion degree of the censoring time distribution, and its calculation formula is: (7) μ c is the mean of censoring time, and its calculation formula is: (8) (2) Calculate the skewness of censoring time s : The censoring time skewness reflects the asymmetry of the censoring time distribution, which is obtained based on the censoring time, the standard deviation, and the mean of the censoring time. Its calculation formula is as follows: (9) S422, based on the censoring rate c and the censoring time skewness s Construct a correction function and use the correction function to correct the mean time to failure (MTTF).

[0047] Assume the estimated value of the initial mean time to failure (MTTF) μ 0≈2 μ c , by introducing the correction function g ( T , K ) adjust the estimated value of the initial mean time to failure (MTTF). For the correction function g ( T , K ), its parameter truncation time T is the average truncation time of the observation interval, representing the average value of the truncation time points of all observation windows, which can reflect the overall length of the observation window. For example, if the components start running at different time points and the end time of each observation window is different, T is the average value of the end time of each observation window, that is, the average value of the running time of the overall components in this stage, which affects the distribution characteristics of the censored data and indirectly affects the censoring rate c and the censoring time skewness s . Another parameter K is the number of data truncation time points, that is, the number of different batches or different observation intervals in the governor system, that is, the number of observation windows, which reflects the complexity of data truncation, K The larger it is, the more observation windows the data is truncated in, and the more complex the censoring mode is, indirectly affecting the censoring rate c . If the components are put into operation in n batches and each batch has an independent observation window, then K = n . Thus, it can be found that the correction function g ( T , K ) can be transformed into a function of the censoring rate c and the censoring time skewness s : (10) where the coefficients a 1, a 2 are determined by multiple linear regression to quantify the impact of censoring characteristics on the estimation error.

[0048] The estimated value of the mean time to failure (MTTF) corrected by the correction function is expressed as: (11) Wherein, μ c is the mean of the censoring times.

[0049] S423. Iteratively optimize the corrected mean time to failure (MTTF): The mean time to failure (MTTF) refers to the average working time of a component before failure, and its calculation formula is: (12) Wherein, Γ ( x ) is the Gamma function, η is the scale parameter, β is the shape parameter, t is the component failure time.

[0050] Step 1: Fix μ 0, substitute the estimated value μ 0 of the mean time to failure (MTTF) corrected by the correction function into Equation (12) to obtain , and substitute it into Equation (5) to estimate the shape parameter β 0 using the maximum likelihood estimation method; Step 2: Fix β 0, use Equation (5) to estimate the scale parameter η* using the maximum likelihood estimation method to obtain , and update the estimated value of the mean time to failure (MTTF) to μ 1 = η* Γ(1 + 1 / β 0); Step 3: Fix μ 1, re-estimate β 1, and determine whether the re-estimated shape parameter satisfies the convergence condition of Equation (13). If not, repeat Steps 1 to 2 until the convergence condition is met.

[0051] (13) Wherein, k is the number of iterations, is the shape parameter iteration threshold, which is set according to needs and actual situations, and is set to 0.01 in this embodiment.

[0052] S430. For an extremely high censoring rate with a censoring rate > 80%, combine the semi-supervised data conversion method and the maximum likelihood estimation method to complete the reliability assessment of key components.

[0053] This method introduces operating condition parameters of the turbine governor components, such as oil pressure, oil temperature, vibration, etc., on the basis of the existing data set, and uses them as the feature vectors of each sample in the data set to establish a data conversion function in the form of Equation (14). The transduction function By comparing the i th fault sample with the j th censored sample j features x i and x j , effectively utilize the information of the fault sample to infer the failure time of the censored sample Infer the pseudo-failure time of the censored sample , use iteration to generate multiple pseudo-candidate times for the censoring time of the censored sample, that is, pseudo-labels, and then select the candidate pseudo-label that minimizes the objective loss function in the form of Equation (15) as the optimal pseudo-label of the censored sample to make it as close as possible to the true failure time T j ; Thus, gradually convert the censored samples into fault samples, and finally use the maximum likelihood estimation method to estimate the parameters for the data set after merging the converted data with the original fault data.

[0054] (14) Wherein: is the pseudo-candidate failure time of the i th fault sample, x i is the feature vector of the i th fault sample, x j is the feature vector of the j th censored sample, C j is the censoring time of the j th censored sample, T i is the true failure time of the i th fault sample, δ j is the censoring index of the j th censored sample, δ j =1 indicates a fault (failure event occurs), δ j =0 indicates censoring (failure event does not occur).

[0055] (15) Formally, a supervised loss function for fault data and an unsupervised loss function for fault or censored data points are given, wherey i is the i th fault sample, is the i pseudo-label of the y j is the j th censored sample, is the j pseudo-label of the W ij contains the edge weights of all node pairs, λ controls the relative importance of the supervision term. The purpose of this function is to select the candidate pseudo-label that minimizes the objective loss function as the optimal pseudo-label for the censored sample.

[0056] The specific steps are as follows: S431, Data preparation: Based on the existing dataset, introduce the operating condition parameters of the turbine governor components such as oil pressure, oil temperature, vibration, etc., and use them as the feature vectors of the dataset samples x . Divide the dataset into two categories, namely fault data and censored data , where x i is the i th feature vector of the fault sample, T i is the i th true failure time of the fault sample, x j is the j th feature vector of the censored sample, C j is the j th censoring time of the observed censored sample.

[0057] S432, Construct the weight matrix by combining feature similarity and survival curve similarity W ij The feature similarity can be obtained by calculating the distance in the feature space using the Gaussian kernel function in Equation (16), and the survival curve similarity can be obtained based on the Kaplan-Meier survival curve difference in Equation (17). For fault samples, the Kaplan-Meier estimator Equation (18) can be used to directly calculate the survival curve. For censored samples, the true failure time is unknown (only known that T j ≥ C j), the global information formula (19) is needed to estimate the survival curve, and the global survival curve is the population average survival function estimated from all samples (including censored samples and failure samples). Then, the combined weights of these two similarities are calculated through formula (18).

[0058] (16) Among them, is the bandwidth parameter, which controls the similarity decay rate. A larger will make the similarity distribution smoother, while a smaller emphasizes the local structure. is the eigenvector x i and x j 's Euclidean distance.

[0059] (17) Among them, controls KS the influence intensity of the difference on the weight, KS is the Kolmogorov - Smirnov statistic, S i ( t ) is the survival function estimate of the i th failure sample.

[0060] (18) Among them, d i is the time t i at which the number of failures occurs, n i is the size of the risk set, indicating the number of samples that have not failed before the time point t i .

[0061] Assume that the survival curve of the censored sample is consistent with the global survival curve after the censoring time C j , that is: (19) (20) Among them, a is the weight coefficient, , which is used to control the weight of the feature similarity.

[0062] S433, calculate the transition matrix based on the weight matrix P : Perform row normalization through formula (19) to ensure that the sum of the transition probabilities of each node is 1, and construct the probability transition matrix P .

[0063] (19) S434. Construct an initial label matrix based on the failure time of the failure samples and the censoring time of the censored samples: The initial label of the failure sample is the failure time , and the initial label of the censored sample is set to the censoring time , and combine the two to obtain the initial label matrix .

[0064] S435. Based on the transition matrix P and the label matrix Y , iteratively update the pseudo-survival time: In a typical right-censoring situation, the actual failure time of the censored sample is unknown but greater than the observed censoring time. Based on this fact, the iteration result is required to enforce the constraint condition after each iteration. In each iteration, the labels of the data points of the censored sample are updated by weighted averaging according to the labels of its neighbor nodes, while fixing the labels of the failure data. The mathematical expression is Equation (20). The transition matrix P and the label matrix Y are partitioned into labeled (subscript l ) and unlabeled (subscript u ) parts as in Equation (21), and the iteration formula can be simplified to update only the unlabeled part as Equation (22). Repeat the above steps until the iteration converges to a convex solution like Equation (23), and the final result is the optimal pseudo-label of the censored sample Y u .

[0065] (22) where is the damping factor, , which is used to control the weight of information propagation.

[0066] (23) (24) (25) S436. Update the dataset After the above steps, the censoring time C j will be replaced by the optimal pseudo-label through the semi-supervised data transformation method, and it is regarded as a failure sample. Update the original dataset: (26) where DFor an updated data set.

[0067] S437, based on the updated data set, the maximum likelihood estimation method is used to complete the reliability assessment of key components. Specifically, assuming that the event time data in the updated data set follows a Weibull distribution, the maximum likelihood estimation is performed using Equation (5) to obtain the final parameter estimation result.

[0068] Based on the same technical concept, an embodiment of the present invention also provides an electronic device, which can implement the reliability assessment method process of the key components of the hydraulic turbine governor provided in the above embodiments of the present invention. In one embodiment, the electronic device can be a server, a terminal device, or other electronic devices. As Figure 2 shown, the electronic device may include: At least one processor, and a memory connected to at least one processor. In the embodiments of the present invention, the specific connection medium between the processor and the memory is not limited. Figure 2 In [the figure], it is taken as an example that the processor and the memory are connected by a bus. The bus is Figure 2 shown by a thick line in [the figure]. The connection manners between other components are only for illustrative purposes and are not to be construed as limiting. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 2 in [the figure] it is only shown by a thick line, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor can also be called a controller, and the name is not limited.

[0069] In the embodiments of the present invention, the memory stores instructions executable by at least one processor. By executing the instructions stored in the memory, at least one processor can execute a reliability assessment method for key components of a hydraulic turbine governor described above. The processor can implement Figure 2 the functions of each module in the device shown in [the figure].

[0070] Among them, the processor is the control center of the device, and can connect various parts of the entire control device through various interfaces and lines. By running or executing the instructions stored in the memory and calling the data stored in the memory, various functions of the device and process data, so as to monitor the device as a whole.

[0071] In an alternative design, the processor may include one or more processing units. The processor may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, and application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above modem processor may not be integrated into the processor. In some embodiments, the processor and the memory can be implemented on the same chip, and in some embodiments, they can also be separately implemented on independent chips.

[0072] The processor can be a general-purpose processor, such as a CPU, a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of a reliability evaluation method for key components of a hydroturbine governor disclosed in the embodiments of the present invention in combination can be directly embodied as being executed by a hardware processor, or executed by a combination of hardware and software modules in the processor.

[0073] As a non-volatile computer-readable storage medium, the memory can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The memory can include at least one type of storage medium, for example, it can include flash memory, hard disk, multimedia card, card-type memory, random access memory (RAM), static random access memory (SRAM), programmable read-only memory (PROM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic memory, magnetic disk, optical disc, and so on. The memory is any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory in the embodiments of the present invention can also be a circuit or any other device capable of implementing a storage function, for storing program instructions and / or data.

[0074] By designing and programming the processor, the code corresponding to the reliability evaluation method for key components of a hydroturbine governor introduced in the foregoing embodiments can be solidified into the chip, so that the chip can execute the steps of the method in the above embodiments when running. How to design and program the processor is a well-known technology to those skilled in the art and will not be elaborated here.

[0075] Based on the same inventive concept, the embodiments of the present invention also provide a storage medium that stores computer instructions, and when the computer instructions run on a computer, the computer is caused to execute a reliability evaluation method for key components of a hydroturbine governor described above.

[0076] In some alternative embodiments, various aspects of a reliability evaluation method for key components of a hydroturbine governor of the present invention can also be implemented in the form of a program product, which includes program code. When the program product runs on a device, the program code is used to cause the control device to execute the steps in a reliability evaluation method for key components of a hydroturbine governor according to various exemplary embodiments of the present invention described above in this specification.

[0077] It should be noted that although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present invention, the features and functions of two or more of the above-described units can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units. Additionally, although the operations of the method of the present invention are described in a specific order in the drawings, this does not require or imply that these operations must be performed in that specific order, or that all of the illustrated operations must be performed to achieve the desired result. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step for execution, and / or one step may be decomposed into multiple steps for execution.

[0078] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0079] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a server, such that the instructions executed by the processor of the computer or other programmable data processing device generate a device for realizing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0080] Program code for performing the operations of the present invention may be written using any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, etc., and also including conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, executed as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0081] In the case of a remote computing device, the remote computing device may be connected to the user's computing device through any type of network including a local area network (LAN) or a wide area network (WAN), or, it may be connected to an external computing device (e.g., by connecting through the Internet using an Internet service provider).

[0082] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device, and the instruction device implements the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 the functions specified in one block or multiple blocks.

[0083] These computer program instructions may also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are performed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 the functions specified in one block or multiple blocks.

[0084] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A reliability evaluation method for key components of a hydraulic turbine governor, characterized in that, Including: Using the analytic hierarchy process for the hydroturbine governor to screen out key components; Processing the original fault operation and maintenance records of the key components to obtain a dataset containing censored data; Calculating the censoring rate based on the dataset; Dynamically selecting the optimal method according to different censoring rates to complete the reliability assessment of key components.

2. The reliability evaluation method for key components of a hydraulic turbine governor according to claim 1, characterized in that The using the analytic hierarchy process for the hydroturbine governor to screen out key components includes: Establishing a hierarchical structure model according to the system structure composition of the hydroturbine governor, and constructing a judgment matrix according to the hierarchical structure model; Using the consistency index to test the consistency of the judgment matrix; After meeting the consistency, screening out key components by calculating the component weights.

3. The reliability evaluation method for the key components of a hydraulic turbine governor according to claim 1, characterized in that The processing the original fault operation and maintenance records of the key components to obtain a dataset containing censored data includes: In the provided original fault operation and maintenance records, screening out the original fault operation and maintenance records of the key components according to the component name and processing them; In the original fault operation and maintenance records of the key components, after normalizing the component commissioning date, until the end of the observation time, the obtained component failure data conforms to the random censoring test, thereby obtaining the number of failed components and the component failure time. The operation time of the components without failures is right censored, thereby obtaining the censoring time; Since the life data of the governor components follows the Weibull distribution, after removing the outliers in the failure time and censoring time, finally obtaining a dataset containing censored data.

4. The reliability evaluation method for key components of a hydraulic turbine governor according to claim 1, characterized in that, The calculating the censoring rate based on the dataset includes: wherein, c is the censoring rate, n is the number of failed components, m is the number of components that have not failed.

5. The reliability evaluation method of the key components of the hydraulic turbine governor according to claim 1, characterized in that, The dynamically selecting the optimal method according to different censoring rate levels to complete the reliability assessment of key components includes: For low censoring rates, using the maximum likelihood estimation method to complete the reliability assessment of key components; For high censoring rates, using a stepwise parameter estimation method that combines censoring characteristic information to complete the reliability assessment of key components; For extremely high censoring rates, combining the semi-supervised data transformation method and the maximum likelihood estimation method to complete the reliability assessment of key components; Among them, the low censoring rate, high censoring rate, and extremely high censoring rate are determined according to a preset censoring rate range.

6. The reliability evaluation method of the key components of a hydraulic turbine governor according to claim 5, characterized in that The using the maximum likelihood estimation method to complete the reliability assessment of key components includes: Among them, L is the likelihood function, n is the number of failed components, m is the number of components that have not failed, t i represents the failure time of the i th failure sample, is the censoring time of the j th censored sample; F ( t ) is the shape parameter of the Weibull distribution β and the scale parameter η of the failure distribution function, expressed as: f ( t ) is the shape parameter of the Weibull distribution β and the scale parameter η of the density distribution function, expressed as: Among them, , t is a random variable, that is, the failure time or censoring time, and exp() is the exponential function with the natural constant e as the base.

7. The reliability evaluation method for the key components of a hydraulic turbine governor according to claim 6, characterized in that, The using a stepwise parameter estimation method that combines censoring characteristic information to complete the reliability assessment of key components includes: Calculating censoring characteristic statistics, including the coefficient of variation of censoring time and the skewness of censoring time; The coefficient of variation of censoring time is the ratio of the standard deviation of censoring time to the mean; It is obtained based on the censoring time and the standard deviation and mean of censoring time; Constructing a correction function based on the censoring rate and the skewness of censoring time, and using the correction function to correct the mean time to failure MTTF; Among them, the correction function is a function based on the censoring rate and the skewness of censoring time; For the corrected mean time to failure MTTF, performing iterative optimization based on the maximum likelihood estimation method: Step 1, fixing the estimated value of the mean time to failure MTTF corrected by the correction function, substituting it into the calculation formula of the mean time to failure MTTF, and substituting the calculated scale parameter into using the maximum likelihood estimation method to estimate the shape parameter; Step 2: Fix the estimated shape parameter, estimate the scale parameter using the maximum likelihood estimation method, and update the estimated value of the mean time to failure (MTTF) using the estimated scale parameter; Step 3: Fix the updated estimated value of the mean time to failure (MTTF), re-estimate the shape parameter according to Step 1, and determine whether the re-estimated shape parameter meets the convergence condition. If it does not meet the condition, repeat Steps 1 to 2 until the convergence condition is satisfied.

8. The reliability evaluation method for the key components of a hydroturbine governor according to claim 7, characterized in that, The reliability assessment of key components completed by combining the semi-supervised data transformation method and the maximum likelihood estimation method includes: Introduce the operating condition parameters of the turbine governor component into the dataset and use them as the feature vectors of the dataset samples; the dataset includes failure samples and censored samples; Construct a weight matrix by combining feature similarity and survival curve similarity; Calculate the transition matrix based on the weight matrix; Construct an initial label matrix based on the failure times of the failure samples and the censoring times of the censored samples; Iteratively update the pseudo-survival time based on the transition matrix and the initial label matrix to obtain the optimal pseudo-labels of the censored samples; Replace the censored samples in the dataset with the optimal pseudo-labels and regard them as failure samples, thereby updating the dataset; Based on the updated dataset, complete the reliability assessment of key components using the maximum likelihood estimation method.

9. An electronic device, characterized in that, It includes: At least one processor; And a memory communicatively connected to the at least one processor; Wherein, the memory stores instructions executable by the at least one processor, and the at least one processor, by executing the instructions stored in the memory, causes the at least one processor to execute the method according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store instructions, and when the instructions are executed, the method according to any one of claims 1-8 is implemented.

Images