Wind turbine reliability evaluation method and system based on multi-level fault system classification

Abstract

The present application relates to the technical field of wind turbine state monitoring and operation and maintenance management, and discloses a wind turbine reliability evaluation method and system based on multi-level fault system classification. The invention establishes a standardized multi-level fault system classification system, attributes the fault outage events of the wind turbine to the subsystems according to the classification system, calculates the multi-dimensional fault evaluation indexes of each subsystem according to the attributed fault outage events, including the single system fault time proportion, the availability damage, the same period change rate and the link period change rate, dynamically sets the threshold value for each dimension of the fault evaluation index of each subsystem, generates and outputs the reliability evaluation visual board of the wind turbine according to the multi-dimensional fault evaluation indexes and the dynamic threshold value of each subsystem, and generates a decision suggestion based on the threshold value rule, which helps personnel quickly locate the weak subsystems and assists personnel in the operation and maintenance management of the unit.

Description

Technical Field

[0001] This invention relates to the field of wind turbine condition monitoring and operation and maintenance management technology, specifically to a wind turbine reliability assessment method and system based on multi-level fault system classification. Background Technology

[0002] Wind turbines typically operate in harsh environments such as sandstorms, salt spray, and alternating loads, leading to frequent failures and high maintenance costs. Currently, wind farm operators generally use macro-level indicators such as "unit average availability" and "total number of failures" to evaluate wind turbine performance and reliability.

[0003] However, the above indicators have significant limitations: First, high availability (e.g., 97%) may mask serious problems such as repeated failures and risk accumulation in critical subsystems. Because macro indicators lump all failures together, they cannot reveal the distribution characteristics of failures across different subsystems, leading to inaccurate judgments by maintenance personnel regarding the true weak points. Second, although the maintenance team is aware of overall turbine failures, it is difficult to accurately pinpoint the specific subsystem causing the longest downtime and most severe losses. Due to the lack of a structured fault attribution system and quantified subsystem-level indicators, the allocation of maintenance resources (such as spare parts and maintenance personnel) lacks focus, resulting in significant blind decision-making. Third, the root cause of power generation loss is difficult to trace back to specific subsystems or components. When negotiating technical improvements or defining responsibilities with equipment suppliers, wind farms can only provide fault statistics at the turbine level, unable to provide precise data on the impact of each subsystem failure on power generation, leading to a lack of objective basis for negotiations and unclear directions for technical improvements. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a wind turbine reliability assessment method based on multi-level fault system classification. This method has the advantages of establishing a standardized multi-level fault system classification system, attributing each fault outage event to subsystems, calculating multi-dimensional fault assessment indicators, and dynamically setting thresholds for each subsystem's assessment indicators to achieve personalized assessment and early warning, thus solving the aforementioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a wind turbine reliability assessment method based on multi-level fault system classification, comprising the following steps: S1: Establish a standardized multi-level fault system classification system; S2: Obtain the fault outage events of the wind turbine units, and perform subsystem attribution for each fault outage event according to the classification system described in S1 to obtain the attributed fault outage events; S3: Based on the attributed outage events, calculate multi-dimensional failure assessment metrics, including the percentage of single-system downtime. Availability impairment Year-on-year change rate and month-on-month change rate ; S4: Establish dynamic thresholds for each subsystem based on attributed failure outage events; S5: Generate and output the reliability assessment results and decision recommendations for wind turbine units based on multi-dimensional fault assessment indicators and dynamic thresholds.

[0006] As a preferred technical solution of the present invention, the S1 multi-level fault system classification system is a tree-like classification map that includes several first-level fault systems, second-level fault components, and third-level fault sub-components, wherein the first-level fault systems, second-level fault components, and third-level fault sub-components are subsystems of the tree-like classification map.

[0007] As a preferred technical solution of the present invention, S2 performs subsystem attribution for each fault outage event, specifically by selecting the last-level node in the tree classification graph through the fault code and recording the start time, end time, and subsystem to which it belongs.

[0008] As a preferred technical solution of the present invention, the percentage of single-system failure time of S3 The expression is as follows:

[0009] in, Indicates the first The percentage of downtime for each subsystem; Indicates the number of times within the statistical period Total downtime of each subsystem; Indicates the total number of subsystems; This represents the total cumulative downtime of all subsystems within the statistical period.

[0010] As a preferred technical solution of the present invention, the S3 availability impairment The expression is as follows:

[0011] in, Indicates the first Availability loss caused by individual subsystems; Indicates the number of times within the statistical period Total downtime of each subsystem; This represents the total calendar time within the statistical period.

[0012] As a preferred technical solution of the present invention, the year-on-year change rate of S3 The expression is as follows:

[0013] in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; Indicates the number of the current statistical period The percentage of downtime for each subsystem; This indicates the number of the previous statistical period in the current statistical period. The percentage of downtime for each subsystem; This indicates a statistical periodic variable.

[0014] As a preferred technical solution of the present invention, the S3 year-on-year change rate The expression is as follows:

[0015] in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; This indicates the number of the current statistical sub-period. The percentage of downtime for each subsystem; This indicates the number of the previous statistical sub-period in the current statistical sub-period. The percentage of downtime for each subsystem; This represents a statistical sub-periodic variable.

[0016] As a preferred embodiment of the present invention, S4 establishes the following expression for the dynamic threshold of each subsystem based on the attributed fault outage events:

[0017] in, Indicates the first Subsystem Dynamic thresholds corresponding to each dimension of fault assessment indicators; This indicates the first historical outage event. Subsystem The average value of each dimension of the fault assessment index; This indicates the first historical outage event. Subsystem Standard deviation of each dimension of fault assessment indicators; Indicates the adjustment coefficient; This indicates the percentage of time spent in a single system failure. Indicates availability impairment. Indicates the year-on-year change rate. This indicates the year-on-year change rate.

[0018] As a preferred technical solution of the present invention, step S5 generates and outputs the reliability assessment results and decision suggestions of the wind turbine based on multi-dimensional fault assessment indicators and dynamic thresholds. Specifically, the calculated indicators are displayed through a visual cockpit dashboard, and decision suggestions are automatically generated based on preset threshold rules.

[0019] As a preferred technical solution of the present invention, the calculated indicators include the percentage of single-system failure time. Availability impairment Year-on-year change rate and month-on-month change rate .

[0020] A wind turbine reliability assessment system based on multi-level fault system classification includes: The classification system module establishes a standardized multi-level fault system classification system; The fault outage event module acquires fault outage events of wind turbine units and performs subsystem attribution for each fault outage event according to the classification system to obtain the attributed fault outage event. The evaluation metrics module calculates multi-dimensional failure evaluation metrics based on the attributed failure and outage events, including the percentage of time spent in a single system failure. Availability impairment Year-on-year change rate and month-on-month change rate ; The dynamic threshold module establishes dynamic thresholds for each subsystem based on the attributed fault outage events; The evaluation module generates and outputs reliability evaluation results and decision recommendations for wind turbines based on multi-dimensional fault assessment indicators and dynamic thresholds.

[0021] Compared with existing technologies, this invention provides a wind turbine reliability assessment method based on multi-level fault system classification, which has the following advantages: This invention establishes a standardized multi-level fault system classification system to attribute wind turbine outage events to subsystems according to the classification system. Based on the attributed outage events, it calculates multi-dimensional fault assessment indicators for each subsystem, including single-system failure time percentage, availability impairment, year-on-year change rate, and month-on-month change rate. It also dynamically sets thresholds for each dimension of the fault assessment indicators for each subsystem. Based on the multi-dimensional fault assessment indicators and dynamic thresholds of each subsystem, it generates and outputs a visual dashboard for the reliability assessment of wind turbines, and generates decision suggestions based on threshold rules to help personnel quickly locate weak subsystems and assist personnel in the operation and maintenance management of the units. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the method of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 A reliability assessment method for wind turbines based on multi-level fault system classification includes the following steps: S1: Establish a standardized multi-level fault system classification system. This system is a tree-like classification diagram that includes several primary fault systems, secondary fault components, and tertiary fault sub-components. Primary fault systems, secondary fault components, and tertiary fault sub-components are subsystems of the tree-like classification diagram. Among them, primary fault systems include transmission and speed change systems, converter systems, generator systems, blade pitch systems, yaw systems, and control systems, etc. The classification diagram is shown in Table 1. Table 1. Tree-like classification diagram (partial)

[0025] S2: Obtain the fault outage events of the wind turbine units, and perform subsystem attribution for each fault outage event according to the classification system of S1 to obtain the attributed fault outage events; For each failure and outage event, subsystem attribution is performed. Specifically, through the human-machine interface, the system receives the last-level node in the tree classification graph selected by the operation and maintenance personnel through the fault code, and records the start time, end time, and subsystem to which it belongs. The specific fault codes are shown in Table 2. Table 2. Common Fault Code Attribution Mapping Table (Partial)

[0026] S3: Based on the attributed outage events, calculate multi-dimensional failure assessment metrics, including the percentage of single-system downtime. Availability impairment Year-on-year change rate and month-on-month change rate ; Percentage of time spent in a single system failure The expression is as follows:

[0027] in, Indicates the first The percentage of downtime for each subsystem; Indicates the number of times within the statistical period Total downtime of each subsystem; Indicates the total number of subsystems; This represents the total cumulative downtime of all subsystems within the statistical period; Traditional wind turbine reliability assessments often use macro-level indicators such as average availability and total number of failures. However, these indicators are insufficient to reveal performance differences between subsystems, easily masking recurring failures and risk accumulation in certain subsystems, and failing to provide the maintenance team with precise identification of weak points. To quantitatively differentiate the contribution of different subsystems to the overall turbine's unreliability from the perspective of downtime, this invention designs a single-system downtime percentage. This indicator, which performs structured attribution statistics on the downtime of each subsystem failure, can intuitively compare which subsystem accounts for the most downtime failures. This indicates that within the statistical period, the [number]th [year]... The percentage of cumulative downtime of a subsystem out of the total downtime of all subsystems; the share of downtime of that subsystem in the total downtime of all subsystems when a wind turbine is shut down due to a fault. The larger the value, the more likely that subsystem is the primary factor causing unit shutdown, i.e., the bottleneck in the overall reliability of the unit; conversely, the smaller the value, the more likely that subsystem is the primary cause. The smaller the value, the smaller the contribution of the subsystem to the unit's failure and shutdown. This indicator directly reflects the relative ranking of the severity of failures in each subsystem and is the core basis for identifying weak links and optimizing the allocation of operation and maintenance resources. Availability Damage The expression is as follows:

[0028] in, Indicates the first Availability loss caused by individual subsystems; Indicates the number of times within the statistical period Total downtime of each subsystem; This represents the total calendar time within the statistical period; In the actual operation and maintenance of wind turbines, the downtime of different subsystems may be the same, but the impact on the overall availability and power generation varies depending on the time of the failure. For example, a failure lasting 10 hours will result in completely different power generation losses and availability impacts depending on whether it occurs during a high-wind season or a low-wind season. The traditional breakdown of downtime percentages... It can only make relative comparisons between subsystems, and cannot perform absolute quantitative loss assessments of the subsystem's downtime within the overall system's operating time scale; to overcome this deficiency, this invention designs availability impairment. The indicator directly correlates the subsystem failure downtime with the total calendar time within the statistical period. It is used to quantify the actual drag on the overall availability of the wind turbine due to the failure of the subsystem, thereby more objectively reflecting the weight of the impact of different subsystems on the overall power generation capacity. Availability Damage This indicates that within the statistical period, the [number]th [year]... The percentage of cumulative downtime due to subsystem failures relative to the total calendar time of the cycle reflects the specific decrease in wind turbine availability directly caused by a failure in that subsystem; if no failure occurs in that subsystem, the turbine availability can be improved accordingly. percentage points; It directly measures the absolute degree of damage to the overall unit's power generation capacity caused by the failure of this subsystem, without being affected by the proportion of failure time in other subsystems; the larger the value of this indicator, the more serious the impact of the failure of this subsystem on the unit's availability and power generation. Year-on-year change rate The expression is as follows:

[0029] in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; Indicates the number of the current statistical period The percentage of downtime for each subsystem; This indicates the number of the previous statistical period in the current statistical period. The percentage of downtime for each subsystem; Indicates a statistical periodic variable; In wind turbine reliability assessment, the percentage of time spent on a single type of failure is... It can only reflect the static relative ranking of each subsystem within the current statistical period, but cannot reveal the dynamic trend of the subsystem's reliability; for example, the current reliability of a certain subsystem... A value of 20% may seem low, but if the same period last year was only 10%, then there is a 100% year-on-year increase, indicating that the performance of this subsystem is rapidly deteriorating and requires early intervention; conversely, if the current... The percentage change is as high as 30%, but it was 40% in the same period last year. Although there has been a decrease, it does not pose an imminent threat. Traditional fixed threshold methods cannot capture such time-varying signals, which can easily lead to missed or false alarms. Therefore, this invention designs a year-on-year change rate. The indicator is obtained by comparing the current statistical period with the previous statistical period of the same length (such as the third quarter of this year with the third quarter of last year). The magnitude of change automatically identifies the deterioration or improvement trends of each subsystem; Indicates the first Failure time percentage of each subsystem The formula represents the percentage change in relative magnitude from the previous statistical period of the same length to the current statistical period. It reflects whether the severity of the subsystem's failure has worsened (positive value) or improved (negative value) compared to the same period last year, and the degree of drastic change. A positive value and a larger absolute value indicate that the proportion of failures in this subsystem is increasing faster year-on-year, and the reliability is showing a significant deterioration trend, even at present. Even a low absolute value should raise concerns; If the value is negative and the absolute value is large, it indicates that the reliability of the subsystem has improved significantly compared to the same period last year. Month-on-month change rate The expression is as follows:

[0030] in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; This indicates the number of the current statistical sub-period. The percentage of downtime for each subsystem; This indicates the number of the previous statistical sub-period in the current statistical sub-period. The percentage of downtime for each subsystem; Represents a statistical sub-periodic variable; Year-on-year change rate While it can reveal the long-term trend of subsystem reliability relative to the same period last year, its time granularity is relatively coarse (usually one year), making it unable to capture rapid fluctuations in the near future (such as adjacent quarters or months). In actual operation and maintenance, the failure performance of some subsystems may deteriorate rapidly in a short period of time. For example, the failure rate of converter systems rises rapidly after entering the high-temperature period of summer. If this is not discovered in the year-on-year analysis, it often results in significant losses. In order to achieve near real-time dynamic monitoring of the reliability status of subsystems, this invention designs a year-on-year change rate. This metric compares the percentage of downtime in the current statistical period (e.g., this quarter) with the percentage in the previous adjacent statistical period (e.g., last quarter). The magnitude of the change can be used to perceive the short-term trend of subsystem reliability at a finer time granularity, providing the operation and maintenance team with timely warnings of deterioration or confirmation of improvement. Indicates the first Failure time percentage of each subsystem The percentage change in relative value from the previous adjacent statistical sub-period to the current statistical sub-period; this index reflects the dynamic direction and severity of the subsystem's fault severity over adjacent short periods. A positive value indicates that the failure rate of the subsystem has increased compared to the previous period, and the reliability has deteriorated recently. The larger the value, the faster the deterioration. A negative value indicates a decrease in the percentage of failures compared to the previous period, suggesting recent improvement in reliability. This indicator can quickly reflect the short-term effects of maintenance measures (such as after a particular overhaul) and can also be used to identify sudden failure modes (such as sudden reliability changes caused by seasonal transitions or changes in operating conditions). (Year-on-year change rate) When used together, they can form a complete perspective on long-term trends and short-term fluctuations; S4: Based on the attributed failure outage events, a dynamic threshold is established for each subsystem, and its expression is as follows:

[0031] in, Indicates the first Subsystem Dynamic thresholds corresponding to each dimension of fault assessment indicators; This indicates the first of the historical outage events. Subsystem The average value of each dimension of the fault assessment index; This indicates the first of the historical outage events. Subsystem Standard deviation of each dimension of fault assessment indicators; Indicates the adjustment coefficient; This indicates the percentage of time spent in a single system failure. Indicates availability impairment. Indicates the year-on-year change rate. Indicates the month-on-month change rate; In wind turbine reliability assessment, if a fixed threshold is used to uniformly judge different subsystems and different periods, a large number of false alarms or missed alarms will occur due to normal fluctuations in indicators caused by factors such as season, turbine age, and wind conditions; for example, the availability of the converter system is affected during the high-temperature period in summer. Naturally high values ​​for this indicator can lead to frequent false alarms if the fixed threshold is too low; conversely, in winter, the normal value for this indicator is lower, and a fixed threshold that is too high may miss genuine abnormal increases. To overcome this drawback, this invention designs a dynamic threshold. The indicator automatically calculates the alarm boundary based on the historical statistical distribution (mean and standard deviation) of indicators in each subsystem and dimension, so that the threshold can adaptively follow the inherent fluctuation characteristics of different subsystems and different indicator dimensions, thereby achieving accurate identification of abnormal changes. Indicates the first The subsystem in the first Dynamic alarm upper bounds across several fault assessment dimensions; among which... This reflects the historical average level (central position) of this dimension indicator of the subsystem. Reflects the historical fluctuation amplitude adjustment coefficient. Used to control the sensitivity of alarms; The physical meaning is: within the current statistical period, if the subsystem's [number]th ... The measured indicators in each dimension exceeded This indicates that the current value of the indicator has significantly deviated from its historical normal fluctuation range, which is a statistically abnormal situation and should trigger an alarm; this dynamic threshold can automatically adapt to the inherent differences of different subsystems (such as the transmission system). The inherently higher (and the different dimensions of various indicators) make the alarm rules uniform and fair, which is the core tool for achieving adaptive, low false alarm rate reliability assessment. S5: Generates and outputs reliability assessment results and decision recommendations for wind turbine units based on multi-dimensional fault assessment indicators and dynamic thresholds. Specifically, it displays the calculated indicators through a visual dashboard and automatically generates decision recommendations based on preset threshold rules. The calculated indicators include the percentage of single-system failure time. Availability impairment Year-on-year change rate and month-on-month change rate ; The threshold rules define the alarm triggering conditions, alarm levels, and text prompts. Alarm levels include red alarm, orange alarm, yellow alarm, and blue prompt. A red alarm occurs when the fault assessment indicators corresponding to three or more dimensions of any subsystem exceed their corresponding dynamic thresholds. An orange alarm occurs when the fault assessment indicators corresponding to two dimensions of any subsystem exceed their corresponding dynamic thresholds. A yellow alarm occurs when the fault assessment indicators corresponding to one dimension of any subsystem exceed their corresponding dynamic thresholds. A blue prompt occurs when the indicator of a certain subsystem of a unit is significantly higher than the average indicator of all subsystems. For example: when the single-system failure time of the converter system accounts for... and year-on-year change rate When all exceed their respective dynamic thresholds, the following message will be displayed: "Orange Alert: The converter system is currently the main reliability bottleneck, and the situation is deteriorating. It is recommended to initiate a deep inspection and spare parts reserve assessment." For example: it is found that the transmission system of Unit A has a single-system failure time percentage. If the value is significantly higher than other units in the same field, the following message will be displayed: "Blue message: There may be an individual defect in the transmission system of Unit A. Targeted condition monitoring is recommended."

[0032] In a specific embodiment of the present invention, a wind turbine reliability assessment system based on multi-level fault system classification includes: The classification system module establishes a standardized multi-level fault system classification system; The fault outage event module acquires fault outage events of wind turbine units and performs subsystem attribution for each fault outage event according to the classification system to obtain the attributed fault outage event. The evaluation metrics module calculates multi-dimensional failure evaluation metrics based on the attributed failure and outage events, including the percentage of time spent in a single system failure. Availability impairment Year-on-year change rate and month-on-month change rate ; The dynamic threshold module establishes dynamic thresholds for each subsystem based on the attributed fault outage events; The evaluation module generates and outputs reliability evaluation results and decision recommendations for wind turbines based on multi-dimensional fault assessment indicators and dynamic thresholds.

[0033] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wind turbine reliability assessment method based on multi-level fault system classification, characterized in that: Includes the following steps: S1: Establish a standardized multi-level fault system classification system; S2: Obtain the fault outage events of the wind turbine units, and perform subsystem attribution for each fault outage event according to the classification system described in S1 to obtain the attributed fault outage events; S3: Based on the attributed outage events, calculate multi-dimensional failure assessment metrics, including the percentage of single-system downtime. Availability impairment Year-on-year change rate and month-on-month change rate ; S4: Establish dynamic thresholds for each subsystem based on attributed failure outage events; S5: Generate and output the reliability assessment results and decision recommendations for wind turbine units based on multi-dimensional fault assessment indicators and dynamic thresholds.

2. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 1, characterized in that: The S1 multi-level fault system classification system is a tree-like classification map that includes several first-level fault systems, second-level fault components, and third-level fault sub-components. The first-level fault systems, second-level fault components, and third-level fault sub-components are subsystems of the tree-like classification map.

3. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 2, characterized in that: S2 performs subsystem attribution for each fault outage event, specifically by selecting the last-level node in the tree classification graph through the fault code and recording the start time, end time, and the subsystem to which it belongs.

4. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 3, characterized in that: The percentage of S3 single system failure time The expression is as follows: in, Indicates the first The percentage of downtime for each subsystem; Indicates the number of times within the statistical period Total downtime of each subsystem; Indicates the total number of subsystems; This represents the total cumulative downtime of all subsystems within the statistical period.

5. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 4, characterized in that: The S3 availability impairment The expression is as follows: in, Indicates the first Availability loss caused by individual subsystems; Indicates the number of times within the statistical period Total downtime of each subsystem; This represents the total calendar time within the statistical period.

6. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 5, characterized in that: The year-on-year change rate of S3 The expression is as follows: in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; Indicates the number of the current statistical period The percentage of downtime for each subsystem; This indicates the number of the previous statistical period in the current statistical period. The percentage of downtime for each subsystem; This indicates a statistical periodic variable.

7. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 6, characterized in that: The S3 year-on-year change rate The expression is as follows: in, Indicates the first The year-on-year change rate of the percentage of single-system failure time in each subsystem; This indicates the number of the current statistical sub-period. The percentage of downtime for each subsystem; This indicates the number of the previous statistical sub-period in the current statistical sub-period. The percentage of downtime for each subsystem; This represents a statistical sub-periodic variable.

8. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 7, characterized in that: The expression for S4 to establish a dynamic threshold for each subsystem based on the attributed fault outage events is as follows: in, Indicates the first Subsystem Dynamic thresholds corresponding to each dimension of fault assessment indicators; This indicates the first historical outage event. Subsystem The average value of each dimension of the fault assessment index; This indicates the first historical outage event. Subsystem Standard deviation of each dimension of fault assessment indicators; Indicates the adjustment coefficient; This indicates the percentage of time spent in a single system failure. Indicates availability impairment. Indicates the year-on-year change rate. This indicates the year-on-year change rate.

9. The wind turbine reliability assessment method based on multi-level fault system classification according to claim 8, characterized in that: The S5 generates and outputs the reliability assessment results and decision suggestions of the wind turbine based on multi-dimensional fault assessment indicators and dynamic thresholds. Specifically, it displays the calculated indicators through a visual cockpit dashboard and automatically generates decision suggestions based on preset threshold rules. The calculated indicators include the percentage of single-system failure time. Availability impairment Year-on-year change rate and month-on-month change rate .

10. A wind turbine reliability assessment system based on multi-level fault system classification, characterized in that, include: The classification system module establishes a standardized multi-level fault system classification system; The fault outage event module acquires fault outage events of wind turbine units and performs subsystem attribution for each fault outage event according to the classification system to obtain the attributed fault outage event. The evaluation metrics module calculates multi-dimensional failure evaluation metrics based on the attributed failure and outage events, including the percentage of time spent in a single system failure. Availability impairment Year-on-year change rate and month-on-month change rate ; The dynamic threshold module establishes dynamic thresholds for each subsystem based on the attributed fault outage events; The evaluation module generates and outputs reliability evaluation results and decision recommendations for wind turbines based on multi-dimensional fault assessment indicators and dynamic thresholds.

Images