Thermal power equipment reliability parameter estimation method and system based on weibull distribution

Abstract

This invention discloses a method and system for estimating reliability parameters of thermal power equipment based on the Weibull distribution, relating to the field of power system equipment reliability assessment technology. The method includes: acquiring multi-source outage event data of the target thermal power equipment during its commissioning period, the multi-source outage event data including at least planned and unplanned outage events; constructing a comprehensive failure time series based on the equipment commissioning time and the occurrence time of each outage event; estimating the shape and scale parameters of the Weibull distribution using the least squares method based on the comprehensive failure time series, obtaining Weibull distribution model parameters adapted to the characteristics of multiple types of outages of thermal power equipment; determining the reliability function of the target thermal power equipment based on the shape and scale parameters, and outputting the reliability assessment results. This invention effectively expands the failure sample size and improves the accuracy and robustness of Weibull distribution parameter estimation by fusing multi-source data, including planned and unplanned outages.

Description

Technical Field

[0001] This invention relates to the field of power system equipment reliability assessment technology, specifically to a method and system for estimating reliability parameters of thermal power equipment based on the Weibull distribution. Background Technology

[0002] In power systems, thermal power equipment operates under complex conditions of high temperature, high pressure, and high load for extended periods, and its reliability directly affects the safe and stable operation of generator units. Wear, bearing failure, and gear failure are common types of failures in thermal power equipment, exhibiting distinct phased characteristics, including early failure phases, intermittent failure phases, and wear-out failure phases. Accurately assessing the reliability of equipment at different operating stages is a crucial foundation for developing preventative maintenance strategies and reducing the risk of unplanned outages.

[0003] Currently, mathematical models used to assess the reliability of mechanical equipment or components include the exponential distribution, normal distribution, log-normal distribution, and Weibull distribution. Among them, the Weibull distribution is widely used in reliability analysis and life assessment due to its good compatibility and strong ability to fit various types of data. However, existing reliability assessment methods based on the Weibull distribution typically rely solely on unplanned outage (failure outage) data for analysis. In practical engineering applications, unplanned outage events in thermal power equipment are relatively few, resulting in insufficient failure sample size and difficulty in accurately fitting the Weibull distribution model parameters, thus affecting the accuracy and robustness of reliability assessment. Furthermore, existing methods often ignore the equipment status information implicit in planned outage events (such as major and minor repairs), failing to fully utilize multi-source outage data throughout the entire life cycle, leading to biased assessment results and an inability to accurately guide the determination of preventive maintenance timing. Summary of the Invention

[0004] To address the problems of insufficient failure sample size and low data utilization leading to poor reliability assessment accuracy in existing technologies, this invention proposes a method and system for estimating reliability parameters of thermal power equipment based on Weibull distribution. By fully utilizing multi-source outage event data and constructing a comprehensive failure time series, it achieves accurate estimation of reliability parameters of thermal power equipment.

[0005] Firstly, a management and control system linking supplier profiles with procurement activities is provided. This method may include: acquiring multi-source outage event data of the target thermal power equipment during its commissioning period, wherein the multi-source outage event data includes at least planned outage events and unplanned outage events; constructing a comprehensive failure time series based on the equipment commissioning time and the occurrence time of each outage event; estimating the shape and scale parameters of the Weibull distribution using the least squares method based on the comprehensive failure time series to obtain Weibull distribution model parameters adapted to the multi-type outage characteristics of thermal power equipment; determining the reliability function of the target thermal power equipment based on the shape and scale parameters, and outputting the reliability assessment results.

[0006] As one implementation method, the planned shutdown events include major overhaul shutdown events and minor overhaul shutdown events.

[0007] As one implementation method, when constructing a comprehensive failure time series, if multiple outage events correspond to the same time point, only the failure hours corresponding to the first event are retained, and the rest are discarded as duplicate values.

[0008] As one implementation method, after estimating the shape and scale parameters of the Weibull distribution using the least squares method, the goodness of fit is used as the evaluation index of model fitness. When the goodness of fit is greater than a preset threshold, the parameter estimation results are confirmed to be valid.

[0009] As one implementation, the reliability assessment result includes mean time between failures (MTBF), which is determined based on the shape parameter, the scale parameter, and the gamma function.

[0010] As one implementation method, when outputting the reliability assessment results, it also includes identifying the running time corresponding to the reliability dropping to a preset threshold, and outputting this running time as a preventive maintenance opportunity.

[0011] In one implementation, the target thermal power equipment includes a coal mill, a feedwater pump set, a blower, an induced draft fan, a high-pressure heater, a desulfurization system, a denitrification system, a circulating water pump, a condensate pump, a primary air fan, an electrostatic precipitator, a bag filter, an electrostatic-bag filter, a wet electrostatic precipitator, a slurry circulating pump, or an oxidation fan.

[0012] As one implementation method, it also includes: using the maximum likelihood estimation method to estimate the parameters of the same comprehensive failure time series, comparing the estimation results of the maximum likelihood estimation method with the estimation results of the least squares method, and outputting the comprehensive parameter estimate when the difference between the two is less than a preset range.

[0013] As one implementation method, the integrated failure time series is sorted in ascending order before parameter estimation.

[0014] In addition, the present invention also provides a reliability parameter estimation system for thermal power equipment based on Weibull distribution, which uses the method described in any of the above-mentioned methods to estimate the reliability parameters of thermal power equipment.

[0015] The present invention provides a method and system for estimating reliability parameters of thermal power equipment based on the Weibull distribution, which has the following beneficial effects: 1. Improved accuracy and sample richness of parameter estimation: This invention overcomes the limitation of traditional methods that rely solely on unplanned outage data by acquiring multi-source outage event data that includes both planned and unplanned outage events. Although planned outage events (such as major and minor repairs) are human-arranged, they reflect, to some extent, the cumulative effect of equipment operation up to that point. Including them in the analysis effectively expands the failure sample size, solving the problem of inaccurate model fitting caused by the scarcity of unplanned outage samples in thermal power equipment, and making the estimation of the shape and scale parameters of the Weibull distribution more robust.

[0016] 2. Enhanced model adaptability to complex operating conditions: By constructing a comprehensive failure time series and employing the least squares method for parameter estimation, the model can comprehensively describe the failure process and characteristics of three different failure stages: early failure, occasional failure, and wear-out failure. In particular, the introduction of goodness-of-fit as an evaluation index ensures the model's fit with actual data and avoids evaluation bias caused by blind modeling.

[0017] 3. Quantitative support for maintenance decisions: This invention constructs a reliability function based on estimated parameters and further calculates the mean time between failures (MTBF) and the operating time when reliability drops to a preset threshold. By quantifying the reliability level of equipment at different operating times, it can accurately pinpoint the time points when reliability rapidly declines, providing a scientific and quantitative basis for formulating preventative maintenance strategies, effectively reducing the risk of unplanned downtime, and improving equipment availability. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for estimating reliability parameters of thermal power equipment based on the Weibull distribution, as provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. In the description of this invention, it should be understood that terms such as "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0022] Example 1: like Figure 1 As shown, this embodiment provides a method for estimating the reliability parameters of thermal power equipment based on the Weibull distribution. This method, by fusing multi-source outage data, solves the model fitting distortion problem caused by the scarcity of unplanned outage samples in traditional methods, achieving accurate estimation of the reliability parameters of thermal power equipment. Specifically, it includes the following steps: Step S100: Obtain multi-source outage event data of the target thermal power equipment during commissioning. The multi-source outage event data includes at least planned outage events and unplanned outage events.

[0023] Specifically, this step aims to overcome the limitations of traditional reliability analysis, which relies solely on unplanned outage (i.e., failure-induced outage) data. In the actual operation of thermal power equipment, unplanned outage events occur relatively infrequently. Using only these as failure samples often results in insufficient sample size, leading to low confidence and large errors in the estimation of Weibull distribution parameters. This embodiment creatively introduces planned outage event data, which can include both major and minor maintenance outages. Although planned outages are not sudden equipment failures, their timing implies the cumulative effect of equipment condition. For example, major maintenance is typically a comprehensive overhaul conducted after a long period of equipment operation and significant wear on critical components. This indicates that while no functional failures have occurred before this point in time, the equipment's reliability level may already be close to a critical state. Therefore, including planned outage events in the analysis sample not only effectively expands the data volume but also fully explores the state evolution information throughout the equipment's entire lifecycle, laying a data foundation for subsequently building a high-precision reliability model.

[0024] Step S200: Starting from the equipment commissioning time, construct a comprehensive failure time series based on the occurrence time of each shutdown event.

[0025] Specifically, after acquiring multi-source outage event data, it needs to be converted into a time series that can be processed by a mathematical model. This embodiment uses the equipment commissioning time (e.g., June 16, 2007) as the origin (zero point) of the time axis, calculating the time difference between the occurrence time of each outage event and the commissioning time; this time difference is the number of failure hours. For example, if an unplanned outage occurs on January 25, 2011, its corresponding failure hours are the difference between that date and the commissioning date. Summarizing the failure hours corresponding to all planned and unplanned outage events forms a comprehensive failure time series. This series no longer distinguishes between event types but uniformly treats them as "state cutoff points" in the equipment operation process, thus constructing a time series reflecting the operational characteristics of the equipment throughout its entire lifecycle.

[0026] Step S300: Based on the comprehensive failure time series, the shape and scale parameters of the Weibull distribution are estimated using the least squares method to obtain the Weibull distribution model parameters that are adapted to the characteristics of multiple types of outages of thermal power equipment.

[0027] Specifically, the Weibull distribution is the core reliability analysis model selected in this embodiment, and its probability density function is determined by the shape parameter β and the scale parameter η. To estimate these two parameters using the integrated failure time series, this embodiment employs the least squares method. Its basic mathematical principle is to linearize the cumulative distribution function of the Weibull distribution. The cumulative distribution function of the Weibull distribution is... By taking two logarithmic transformations, the formula can be transformed into a linear equation. ,in In the specific calculation, the comprehensive failure time series is first sorted, and the cumulative failure probability F(t) corresponding to each failure time point is estimated using the median rank method, thus obtaining a series of (X_i, Y_i) data points. Then, using the least squares principle, an optimal fitting line is found that minimizes the sum of the squared distances from all data points to this line. The shape parameter β and the scale parameter η can be derived from the slope and intercept of the fitted line. The shape parameter β reflects the failure mode of the equipment; for example, when β is greater than 1, it indicates that the equipment is in the wear-out failure period, and the failure rate accelerates with time. The scale parameter η reflects the characteristic life of the equipment.

[0028] Step S400: Determine the reliability function of the target thermal power equipment based on the shape parameters and dimensional parameters, and output the reliability assessment results.

[0029] Specifically, after determining the shape parameter β and the scale parameter η, the reliability function of the target thermal power equipment can be constructed. This function describes the probability that the equipment will still be able to perform its intended functions normally after an operating time t. The reliability assessment results not only include this function expression, but can also further include quantitative indicators such as reliability values ​​at key time points and mean time between failures (MTBF). By outputting these assessment results, maintenance personnel can intuitively understand the current reliability status of the equipment and its future trends, thus providing data support for developing scientific maintenance strategies. For example, if the calculation results show that the equipment's reliability drops sharply after 30,000 hours of operation, preventative maintenance can be scheduled in advance to avoid unplanned downtime.

[0030] Example 2: Based on Example 1, this example refines the process of classifying multi-source outage event data and constructing a comprehensive failure time series.

[0031] Specifically, planned outage events can include major overhaul outages and minor overhauls. Major overhaul outages typically refer to outages involving comprehensive disassembly and inspection, replacement of major components, or complete machine repair. Their occurrence often signifies that the equipment has undergone a long operating cycle, and the accumulated wear and tear has reached a point requiring full restoration. Minor overhaul outages, on the other hand, refer to repairs targeting localized defects or specific components, with a relatively short cycle. Although these two types of events are preventative measures arranged by the operator, their occurrence time objectively records the "survival time" of the equipment up to that point without functional failure. Incorporating major and minor overhaul events into the failure time series effectively utilizes the equipment's state information throughout its entire lifecycle, solving the sample sparsity problem caused by solely relying on unplanned outage data, thereby improving the robustness of parameter estimation.

[0032] When constructing a comprehensive failure time series, if multiple outage events correspond to the same time point, only the failure hours corresponding to the first event are retained, and the rest are discarded as duplicates. In actual engineering data recording, due to system input delays or manual operation habits, multiple outage events of different types may be recorded at the same time. For example, equipment may first experience an unplanned outage and then proceed to a major overhaul process, resulting in two records, "unplanned outage" and "major overhaul outage," existing simultaneously at the same timestamp in the database. In this case, this embodiment sets a priority rule, prioritizing the retention of unplanned outage events because they directly represent the actual physical moment of equipment failure, while planned outages are often subsequent response measures. By removing duplicates, the uniqueness of each time point in the time series is ensured, avoiding errors in failure probability calculation caused by duplicate counting and guaranteeing the accuracy of the Weibull distribution fitting.

[0033] Furthermore, the composite failure time series is sorted in ascending order before parameter estimation. Since the acquired outage event data is typically arranged chronologically rather than by failure hours, directly using the raw sequence would lead to logical inconsistencies in subsequent parameter estimation. The ascending-ordered sequence accurately corresponds to the cumulative failure probability; that is, the earlier the failure time, the lower the corresponding cumulative failure probability. This preprocessing step not only satisfies the mathematical prerequisites for calculating the cumulative distribution function using the least squares or median-rank method but also significantly improves the computational efficiency and convergence speed of subsequent parameter estimation.

[0034] Example 3: Based on Example 1 or Example 2, this example further optimizes and verifies the accuracy and robustness of the Weibull distribution parameter estimation.

[0035] Specifically, after estimating the shape and scale parameters of the Weibull distribution using the least squares method, the goodness of fit is used as an evaluation index for model fitness. When the goodness of fit is greater than a preset threshold, the parameter estimation results are considered valid. Goodness of fit is a key indicator for measuring the degree of fit between the Weibull distribution model and actual failure data, and it is typically characterized by the correlation coefficient Rho. In the linearized coordinate system of Weibull probability paper, the closer the Rho value is to 1, the stronger the linear correlation between the data points and the fitted line, and the more accurate the model's description of the failure pattern. In this embodiment, the preset threshold can be set to 0.8. If the calculated Rho value is greater than 0.8, it is determined that the current comprehensive failure time series has good consistency with the Weibull distribution model, and the estimated shape parameter β and scale parameter η can truly reflect the equipment failure pattern. If the Rho value is lower than this threshold, it indicates that there may be outliers in the data or that the equipment failure mode has changed (such as a mixed failure mode). At this time, the system can issue an early warning, prompting maintenance personnel to check the data quality or try other distribution models. The introduction of this verification step effectively avoids forcibly outputting incorrect parameters when the data fit is poor, thus ensuring the credibility of the subsequent reliability assessment results.

[0036] As a preferred implementation, this embodiment further includes: estimating parameters for the same comprehensive failure time series using the maximum likelihood estimation method, and comparing the estimation results of the maximum likelihood estimation method with those of the least squares method. When the difference between the two is less than a preset range, the comprehensive parameter estimate is output. Although the least squares method is computationally simple and intuitive for fitting linearized data, its statistical characteristics are not as good as the maximum likelihood estimation method (MLE) when dealing with truncated data or small sample sizes. The MLE method constructs a likelihood function based on the probability density function and obtains parameter estimates by solving the extreme points of the log-likelihood equation system, exhibiting asymptotic unbiasedness and asymptotic efficiency. This embodiment constructs a more rigorous parameter estimation logic through dual algorithm cross-validation. The specific comparison process can be set as follows: calculating the relative deviation between the shape parameter β_LSE and β_MLE obtained by the two methods, i.e., |β_LSE-β_MLE| / β_LSE. If the relative deviation is less than a preset range (e.g., 5% or 10%), it indicates that the two algorithms are statistically consistent, and the parameter estimation results are robust and reliable. At this point, the system can output a comprehensive parameter estimate, such as the arithmetic mean or weighted average of the two (the weights can be determined based on the goodness of fit), to further eliminate random errors from a single algorithm. If the difference between the two exceeds a preset range, it indicates that the data may have special structure or outliers, and the system can mark the evaluation result for manual review. This technique of integrating multiple algorithms significantly improves the anti-interference capability of reliability parameter estimation for thermal power equipment, prevents the distortion of evaluation results due to the limitations of a single algorithm, and provides a solid data foundation for formulating scientific maintenance strategies.

[0037] Example 4: Based on any one of Examples 1 to 3, this example describes in detail the quantitative output of reliability assessment results and their application in maintenance decisions, transforming the abstract mathematical model into quantitative indicators that can guide engineering practice.

[0038] Specifically, the reliability assessment results may include Mean Time Between Failures (MTBF), which is determined based on the shape parameter, the scale parameter, and the gamma function. MTBF is a core indicator for measuring equipment reliability, representing the average uptime of the equipment between two failures. In this embodiment, the formula for calculating MTBF is as follows: ; Where η is the scale parameter of the Weibull distribution, β is the shape parameter, and Γ( () is the gamma function. The gamma function is defined as follows: In practical engineering calculations, the value of MTBF is usually obtained by consulting a gamma function table or using a numerical integration algorithm. For example, for the desulfurization system of Unit 8 at the Dalate Power Plant, if the estimated shape parameter β = 2.0, then 1 + 1 / β = 1.5. By consulting a table, we know that Γ(1.5) = π / 2 ≈ 0.88623. Substituting the scale parameter η = 38260 hours, the calculated MTBF is approximately 33910 hours. This quantitative indicator directly reflects the mean time between failures (MTBF) of the equipment under current operating conditions, providing a benchmark reference for maintenance personnel to assess the health status of the equipment. It should be understood that although the above formula provides the standard calculation method for MTBF, in scenarios involving the fusion of multi-source outage data, in order to further eliminate the non-fault interference of planned outage events on the statistics of fault interval time, in a preferred embodiment, an event type correction coefficient can be introduced to correct the MTBF, making it closer to the actual fault interval level of the equipment, thereby improving the engineering practicality of the assessment results.

[0039] Furthermore, when outputting the reliability assessment results, the system also includes identifying the runtime corresponding to when the reliability drops to a preset threshold, and outputting this runtime as the timing for preventative maintenance. This process represents a leap from "condition assessment" to "decision support." The reliability function of the Weibull distribution... It is a function that monotonically decreases with time t. This embodiment sets a preset reliability threshold. (e.g., 0.6, 0.5, or 0.4), and then solve for the corresponding running time. This is a suggested timing for preventative maintenance. The specific calculation formula is as follows: Where η is the scale parameter and β is the shape parameter. Preset threshold. The settings are not fixed but dynamically adjusted based on the importance of the target thermal power equipment in the power generation system, the risk of downtime losses, and the operation and maintenance costs. For example, the shutdown of key auxiliary equipment such as coal mills and feedwater pump sets may lead to load reduction or even unscheduled shutdowns of the unit; therefore, the settings can be adjusted accordingly. A higher threshold (e.g., 0.7 or 0.8) can be set for early warning; for redundant equipment, the threshold can be appropriately lowered. Taking the aforementioned desulfurization system as an example, if a reliability threshold is set... The value is 0.6. Substituting the parameters, the calculation yields 27,344 hours, meaning that the reliability will drop to 60% after approximately 3.86 years of operation. At this point, the system automatically outputs a "recommended maintenance" signal. Maintenance personnel can then arrange preventative maintenance based on this signal, effectively preventing the equipment from entering a high-failure-rate zone (wear-out failure period). This transforms traditional "reactive repairs" or blind "periodic maintenance" into precise "condition-based maintenance," significantly reducing the risk of unplanned outages and improving the economic efficiency and safety level of thermal power plants.

[0040] Example 5: This embodiment uses a specific application scenario as an example to verify the above method flow in detail. The target thermal power equipment can specifically be a coal mill, feedwater pump set, forced draft fan, induced draft fan, high-pressure heater, desulfurization system, denitrification system, circulating water pump, condensate pump, primary air fan, electrostatic precipitator, bag filter, electrostatic-bag filter, wet electrostatic precipitator, slurry circulation pump, or oxidation fan. It should be understood that the above equipment is only illustrative, and the method provided by this invention is applicable to all rotating machinery or systems in thermal power plants with a clearly defined commissioning time and a record of shutdown events.

[0041] Specifically, the desulfurization system of Unit 8 at the Dalate Power Plant was selected as the evaluation object. This unit was put into operation on June 16, 2007, and the statistical cutoff date is June 2024. In step S100, the system acquires multi-source outage event data for the desulfurization system during its operation. A total of 21 outage events were obtained, including 13 unplanned outage events (UO), 2 major overhaul outage events (PO1), and 6 minor overhaul outage events (PO2). If traditional methods are used, only 13 unplanned outage events are extracted as failure samples, resulting in a small sample size and making it difficult to construct a high-confidence Weibull distribution model. This embodiment includes both major and minor overhaul outage events in the analysis, effectively expanding the sample size.

[0042] In step S200, starting from the equipment commissioning date of June 16, 2007, the time difference between the start time and commissioning time of each shutdown event is calculated to obtain the failure hours. After deduplication of duplicate events at the same time point, the constructed comprehensive failure time series (unit: hours) is as follows: 31671, 31761, 31956, 32386, 32457, 32466, 33597, 33644, 33956, 34671, 34965, 36191, 40060, 40728, 48142, 73069, 89684, 100152, 117046, 137832, 145945. This series covers the entire life cycle state change nodes of the equipment from the initial commissioning to the end of operation.

[0043] In step S300, based on the aforementioned comprehensive failure time series, the least squares method is used to estimate the Weibull distribution parameters. The calculated shape parameter β is 20.6787, and the scale parameter η is 34642.13 hours. Simultaneously, the goodness-of-fit Rho value is calculated to be 0.8414. This goodness-of-fit is greater than a preset threshold (e.g., 0.8), indicating that even after including planned outage events in the analysis, the data still highly conforms to the Weibull distribution law, proving the effectiveness of the multi-source data fusion strategy of this invention. If only 13 unplanned outage data points are used for fitting, the linearity of the fitted line is often poor due to the small sample size, making it difficult to accurately determine the failure mode. This embodiment, by introducing planned outage data, allows the shape parameter β to more robustly reflect the characteristics of the equipment being in its wear-out and failure period (β is much greater than 1).

[0044] In step S400, a reliability function is determined based on the estimated parameters, and the evaluation results are output. The calculated mean time between failures (MTBF) is approximately 34,280 hours (approximately 3.91 years). The reliability evaluation results show that the reliability of the desulfurization system is approximately 0.92 after 30,000 hours of operation, decreasing to approximately 0.58 after 35,000 hours, and only 0.03 after 40,000 hours. This data trend indicates that the equipment's reliability begins to decline rapidly after approximately 3.5 years of operation. Based on this, the system identifies the operating time corresponding to a reliability drop to a preset threshold (e.g., 0.6) of approximately 33,800 hours and outputs this as a preventative maintenance opportunity. Maintenance personnel can then schedule preventative maintenance approximately 3.86 years after the equipment's operation to avoid the equipment entering a high failure rate zone and reduce the risk of unplanned downtime.

[0045] Furthermore, the method provided by this invention is also applicable to other key auxiliary equipment in Unit 8 of the Dalate Power Plant, such as the coal mill, feedwater pump set, forced draft fan, induced draft fan, and high-pressure heater. These devices also exhibit the characteristic of having few unplanned outage samples but abundant planned outage records. Utilizing the method of this invention can fully leverage the value of the entire lifecycle data and improve the accuracy of reliability assessment.

[0046] Example 6: This embodiment provides a reliability parameter estimation system for thermal power equipment based on the Weibull distribution. The system is used to execute the method described in any one of embodiments 1 to 5 above, and realizes the automated assessment of the reliability of thermal power equipment through a combination of software and hardware.

[0047] Specifically, the overall architecture of the system mainly includes a data acquisition module, a sequence construction module, a parameter estimation module, and a result output module. It should be understood that the above module division is only a logical functional division. In actual implementation, these modules can be stored in a computer-readable storage medium for processor execution, or some or all functions can be implemented using hardware circuits (such as FPGAs, ASICs, etc.).

[0048] The data acquisition module is used to acquire multi-source outage event data of the target thermal power equipment during commissioning. This multi-source outage event data includes at least planned and unplanned outage events. As the system's input, this module typically communicates with the power plant's existing reliability management system, DCS system, or MIS system via a data interface. Specifically, the data acquisition module is equipped with a standardized data cleaning interface capable of identifying and extracting key fields such as equipment commissioning time, event start time, and event status type (e.g., UO, PO1, PO2). At the hardware level, this module can be implemented as a network interface card or a serial communication interface, reading raw operating logs from an external database via Modbus, OPC, or TCP / IP protocols, thus solving the problems of low efficiency and error-proneness associated with traditional manual data entry.

[0049] The sequence construction module is used to construct a comprehensive failure time series based on the occurrence time of each shutdown event, starting from the equipment commissioning time. This module integrates a time difference calculation unit and a deduplication and sorting unit. Specifically, the time difference calculation unit reads the shutdown event records output by the data acquisition module, calculates the time difference between the event start time and the equipment commissioning time, and obtains the number of failure hours in hours. The deduplication and sorting unit executes the deduplication logic described in the previous embodiment; that is, when multiple records corresponding to the same timestamp are detected, a unique record is retained according to a preset event priority (e.g., unplanned shutdown takes precedence over planned shutdown), and all failure hours are sorted in ascending order, finally outputting a standardized comprehensive failure time series. This module's design transforms discrete, multi-source event records into a numerical sequence that can be directly processed by a mathematical model, providing standardized data input for subsequent parameter estimation.

[0050] The parameter estimation module is used to estimate the shape and scale parameters of the Weibull distribution based on the comprehensive failure time series using the least squares method, thereby obtaining Weibull distribution model parameters adapted to the characteristics of various types of outages in thermal power equipment. This module is the core computing unit of the system, and it has a built-in Weibull distribution linearization processing algorithm and a least squares solver. Specifically, the parameter estimation module receives the time series output by the sequence construction module, first calculates the cumulative failure probability corresponding to each time point using the median rank method, then performs a double logarithmic transformation to construct a linear regression model, and finally solves for the shape parameter β and scale parameter η using the least squares method. In a preferred embodiment, this module also integrates a verification submodule for calculating the goodness of fit Rho. The parameters are only considered valid when the Rho value is greater than a preset threshold (e.g., 0.8); otherwise, an alarm is triggered. In addition, this module can optionally be configured with a maximum likelihood estimation (MLE) algorithm unit for cross-validation with the aforementioned least squares results, further ensuring the robustness of the parameter estimation.

[0051] The result output module is used to determine the reliability function of the target thermal power equipment based on the shape and scale parameters, and output the reliability assessment results. This module is responsible for transforming abstract mathematical parameters into intuitive engineering indicators. Specifically, the result output module has a built-in gamma function calculation unit, used to calculate the mean time between failures (MTBF) based on β and η. Simultaneously, this module is also equipped with an inverse solution unit, capable of inversely solving for the corresponding preventative maintenance timing based on a preset reliability threshold (e.g., 0.6). The output format can be a visual chart (e.g., a curve showing reliability changes over time), or a specific numerical report or maintenance recommendation report. At the hardware level, this module can connect to the power plant's display terminal or printer to intuitively display the assessment results to maintenance personnel, thereby realizing a fully automated closed loop from data acquisition, processing, analysis to decision support.

[0052] Through the design of the above system architecture, the previously fragmented calculation process, which relied on manual experience or single software tools, has been integrated into a standardized automatic evaluation system. The smooth data flow and tight logical coupling between modules not only significantly improve the efficiency and accuracy of reliability assessment for thermal power equipment, but also effectively prevent competitors from circumventing the scope of protection of this invention by developing similar software systems, providing strong technical support for the intelligent operation and maintenance of thermal power plants.

[0053] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention, such as equivalent substitutions for parameter estimation algorithms, specific numerical adjustments to the goodness-of-fit threshold, or extended applications to target device types, should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for estimating reliability parameters of thermal power equipment based on Weibull distribution, characterized in that, The method includes: Acquire multi-source outage event data of the target thermal power equipment during commissioning, wherein the multi-source outage event data includes at least planned outage events and unplanned outage events; Starting from the equipment commissioning time, a comprehensive failure time series is constructed based on the occurrence time of each shutdown event; Based on the comprehensive failure time series, the shape and scale parameters of the Weibull distribution are estimated using the least squares method to obtain Weibull distribution model parameters that are adapted to the characteristics of multiple types of outages of thermal power equipment. The reliability function of the target thermal power equipment is determined based on the shape parameters and dimensional parameters, and the reliability assessment result is output.

2. The Weibull distribution-based reliability parameter estimation method for thermal power plants according to claim 1, characterized by, The planned outage events include major overhaul outage events and minor overhaul outage events.

3. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, When constructing a comprehensive failure time series, if multiple outage events correspond to the same time point, only the failure hours corresponding to the first event are retained, and the rest are discarded as duplicate values.

4. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, After estimating the shape and scale parameters of the Weibull distribution using the least squares method, the goodness of fit is used as the evaluation index of model fitness. When the goodness of fit is greater than the preset threshold, the parameter estimation results are confirmed to be valid.

5. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, The reliability assessment results include mean time between failures (MTBF), which is determined based on the shape parameter, the scale parameter, and the gamma function.

6. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, When outputting reliability assessment results, the system also includes identifying the running time corresponding to when the reliability drops to a preset threshold, and outputting this running time as the timing for preventive maintenance.

7. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, The target thermal power equipment includes coal mills, feedwater pump sets, blowers, induced draft fans, high-pressure heaters, desulfurization systems, denitrification systems, circulating water pumps, condensate pumps, primary air fans, electrostatic precipitators, bag filters, electrostatic-bag filters, wet electrostatic precipitators, slurry circulating pumps, or oxidation fans.

8. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, It also includes: using the maximum likelihood estimation method to estimate the parameters of the same comprehensive failure time series, comparing the estimation results of the maximum likelihood estimation method with the estimation results of the least squares method, and outputting the comprehensive parameter estimate when the difference between the two is less than a preset range.

9. The method for estimating reliability parameters of thermal power equipment based on Weibull distribution according to claim 1, characterized in that, The comprehensive failure time series is sorted in ascending order before parameter estimation.

10. A reliability parameter estimation system for thermal power equipment based on Weibull distribution, characterized in that, The reliability parameters of thermal power equipment are estimated using the method described in any one of claims 1 to 9.

Images