A comprehensive reliability assessment method and related apparatus for DC wind farms

By establishing a power output state model and a fault tree model for DC wind farms, and combining this with a normal cloud generator to calculate cloud droplet impact values, a multi-level reliability index system is formed. This solves the problems of unclear index system and unreasonable weight calculation in the reliability assessment of DC wind farms, and realizes multi-dimensional reliability assessment and technical support.

CN123121503APending Publication Date: 2026-11-17XINJIANG UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611306852.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing reliability assessment methods for DC wind farms suffer from unclear logic in the construction of the indicator system and a lack of rationality in the weight calculation, making it difficult to fully reflect the multi-dimensional reliability characteristics of DC wind farms.

Method used

A power output state model is established using DC wind farm topology, component reliability data, and historical wind speed data. Logical judgment is performed using a fault tree model. Cloud droplets are generated using a normal cloud generator, and the average impact value of cloud droplets is calculated to form a multi-level reliability index system. A comprehensive reliability evaluation index is then calculated.

Benefits of technology

It enables a comprehensive assessment of the multi-dimensional reliability of DC wind farms, improving the rationality and accuracy of the assessment results and providing technical support for wind farm planning, operation and maintenance, and grid connection of new energy sources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121503A_ABST
    Figure CN123121503A_ABST
Patent Text Reader

Abstract

This invention discloses a comprehensive reliability assessment method and related apparatus for DC wind farms, belonging to the technical field of comprehensive reliability assessment for wind farms. The method includes: establishing a DC wind turbine output state model based on historical wind speed data, and establishing a component outage probability model based on component reliability data; establishing a DC wind farm fault tree model based on the DC wind farm topology; defining reliability assessment indicators based on the total output power after turbine disconnection, forming a multi-level reliability indicator system for the DC wind farm; obtaining the weights of the corresponding reliability indicators through a cloud droplet integrated cloud model weighting method; and calculating the comprehensive reliability assessment index of the DC wind farm based on each reliability indicator and its weight, as the comprehensive reliability assessment result of the DC wind farm. This invention achieves a comprehensive assessment of the multi-dimensional reliability of DC wind farms, providing technical support for the planning, operation, and maintenance of DC wind farms and the grid connection of new energy sources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the technical field of comprehensive reliability assessment of wind farms, and relates to a method and related apparatus for comprehensive reliability assessment of DC wind farms. Background Technology

[0002] Currently, wind power development is showing strong growth momentum. Researching the reliable collection and transmission capabilities of wind farms has become a key issue in ensuring the efficient utilization of new energy sources and promoting the transformation of the energy structure. Considering that large-scale wind power can fully utilize the technical advantages of DC transmission when connected to the AC grid via DC transmission lines, effectively improving the efficient access capability of new energy sources, the DC wind farm (DC collection-DC transmission) method, with its unique advantages of low loss and high efficiency, effectively overcomes problems such as harmonic resonance and reactive power transmission in traditional AC collection systems, and has gradually gained widespread attention from the industry.

[0003] Numerous achievements have been made in the research of wind farm reliability assessment. Existing technologies use the Expected Energy Shortage Value (EENS) to assess wind farm reliability, or use the Probability of Load Loss (LOLP), Time of Load Loss (LOLE), and EENS to describe the reliability of wind farms from the perspectives of probability, duration, and energy loss, respectively. However, DC wind farms have dispersed turbine layouts and are spatially distant from load centers. Their reliability assessment requires the construction of a multi-dimensional assessment system covering the entire chain from the generator end to the DC collection link, equipment end, transmission link, and power receiving end. This system must consider spatial dimension indicators such as the collection and transmission links, as well as temporal dimension indicators such as equipment failure cycles. At the same time, it is necessary to establish an energy conversion efficiency constraint model that includes DC line impedance and power semiconductor device conduction losses to comprehensively reflect the reliability characteristics of the coordinated operation of DC wind farms.

[0004] In terms of comprehensive evaluation, the improved entropy weighting method may lead to weight allocation relying too heavily on data distribution and neglecting the interrelationships between indicators. While the improved DEMATEL-AHP-EWM method considers the influence of subjective judgment and objective data on indicator weights, it fails to account for the fact that reliability indicators, while possessing a certain degree of randomness and difficult to measure precisely, may also be affected by various random factors. Traditional cloud model weighting methods rely solely on three numerical characteristics of the cloud model (expectation, entropy, and hyperentropy) to calculate Hamming distance to determine weights, failing to fully consider the inherent randomness of cloud models and neglecting the impact of cloud droplet location distribution on weights. Summary of the Invention

[0005] The purpose of this invention is to provide a comprehensive reliability assessment method and related apparatus for DC wind farms, in order to solve the problems of existing comprehensive reliability assessment methods in terms of unclear logic in the construction of the indicator system and lack of rationality in weight calculation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a comprehensive reliability assessment method for DC wind farms, comprising: Acquire the topology, component reliability data, and historical wind speed data of the DC wind farm; establish a DC wind turbine output state model based on the historical wind speed data, and establish a component outage probability model based on the component reliability data; randomly sample the output state of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output state model, and calculate the total output power of the DC wind farm. A fault tree model of a DC wind farm is established based on the topology of the DC wind farm. Logical judgments are made on the state of each system based on the fault tree model to analyze whether the wind turbines need to be disconnected and the number of wind turbines to be disconnected. The total output power of the remaining normally operating wind turbines in the DC wind farm after the disconnection is calculated. Based on the total output power after the turbine is cut off and the DC loss of the DC wind farm, reliability evaluation indicators are defined from the aspects of space and object, and a multi-level reliability index system for DC wind farms is calculated and formed. The cloud droplets corresponding to each reliability index are generated using a normal cloud generator. Based on the distance between the cloud droplets and the desired outcome and the membership degree of the cloud droplets, the average influence value of the cloud droplets corresponding to each index is calculated, and the average influence value of the cloud droplets of each cloud model is normalized to obtain the weight of the corresponding reliability index. Based on each reliability index and its weight, a comprehensive reliability evaluation index for DC wind farms is calculated, which serves as the comprehensive reliability evaluation result for DC wind farms.

[0007] Furthermore, the acquisition of DC wind farm topology, component reliability data, and historical wind speed data; and the establishment of a DC wind turbine output state model based on the historical wind speed data, includes: The wind speed model is constructed using the Weibull distribution, whose specific probability density function and cumulative distribution function are expressed as follows:

[0008] In the formula, The shape parameter of the Weibull distribution can characterize the range of wind speed variation. The larger the value, the smaller the distribution range of wind speed; is the scale parameter of the Weibull distribution; F is a random number that follows a uniform distribution in [0,1]. The conversion relationship between fan output power and wind speed is as follows:

[0009] In the formula, v ci To cut in wind speed; vco To cut off the wind speed; v R Rated wind speed; P R This refers to the rated power of a single fan. K-means clustering was used to classify and analyze wind power data over a certain time period. K-means iteratively updates cluster centers and redistributes data points, ultimately dividing the dataset into K clusters. Data points within each cluster are relatively close in the feature space, while data points in different clusters are relatively far apart. The specific process is as follows: (1) Input all wind turbine output datasets Assuming the maximum number of clusters is Kmax, and the initial number of clusters is k (k≤Kmax), the output data is divided into... There are k clusters in total, and k initial points are randomly selected. As the initial cluster center; (2) Calculate the output data With each cluster center Manhattan distance between them:

[0010] (3) Calculate After determining the distances to the k cluster centers, Assign it to the cluster corresponding to the nearest cluster center; (4) Recalculate the cluster centers based on the output data contained in each cluster after the update:

[0011] In the formula, For the updated cluster The new cluster center; For the updated cluster The number of output data; (5) Repeat steps (2) to (4) until the change in the cluster center value of each cluster is less than the set threshold or the set maximum number of iterations is reached; (6) Calculate the average silhouette coefficient, which ranges from [-1, 1]. The closer it is to 1, the better the clustering effect. Its expression is:

[0012] In the formula, The mean silhouette coefficient is the number of data clusters divided into k clusters. for The average distance to other data in the same cluster. for The average distance to all data in the nearest cluster; (7) Compare the size of SC under different k values, select the k value corresponding to the maximum value of SC as the optimal number of clusters K, and count the number of output data and cluster center value contained in each cluster. (8) Output the cluster centers cj of the K clusters and their state probabilities. ,in Divide the number of output data points in each cluster by the total number of output data points, i.e.:

[0013] Furthermore, the component failure probability model is established by combining component reliability data; the output status of the DC wind turbines in the system is randomly sampled, and the output of each wind turbine is determined by combining the DC wind turbine output status model, and the total output power of the DC wind farm is calculated, including: Input data related to the topology and component reliability of the DC wind farm; A power output state model for DC wind turbines was established based on historical wind speed data, and a component downtime probability model was established in conjunction with equipment reliability parameters. The output status of the DC wind turbines in the system is randomly sampled. Based on the output status model results of the DC wind turbines, the output of each wind turbine is determined, and the total output power of the DC wind farm in this sampling is calculated.

[0014] Furthermore, a fault tree model of the DC wind farm is established by combining the topology of the DC wind farm; based on the fault tree model, logical judgments are made on the state of each system to establish a fault tree model of the DC wind farm. The sequential Monte Carlo simulation method was used to sample the operating states of each component in the DC wind farm, and the time-series state transition sequence of each component and the overall time-series state transition sequence of the system were obtained. Based on the aforementioned fault tree model of DC wind farm, the system state is logically judged in sequence, and it is analyzed whether the system needs to disconnect the turbines and the number of turbines to be disconnected under the state. Calculate the total output power of the remaining normally operating wind turbines in the DC wind farm after the turbines are switched off. Calculate the reliability indicators of a DC wind farm; Determine whether the reliability index meets the set calculation period. If it does, output the reliability index of the DC wind farm; otherwise, return.

[0015] Furthermore, the step of basing the total output power after turbine disconnection and the DC loss of the DC wind farm includes: DC fan power loss:

[0016] In the formula, P g,loss For generator losses; Psw,vsc For converter switching losses; P c,vsc This refers to the converter's conduction loss. f s This refers to the converter switching frequency; E von,I , E on,I These are the IGBT turn-on losses and turn-off losses within the converter, respectively. E off,D This refers to the diode turn-off loss within the converter; U v,ref This is the rated voltage of the converter; I v,ref This is the rated current of the converter; I wt This represents the output current of a single fan. U wt This refers to the output voltage of a single fan. N vsc This refers to the number of IGBTs in the converter; P vcond,I , P vcond,D These are the conduction losses of the IGBTs and diodes within the converter, respectively. DC / DC converter power loss: The on-state loss PCL, I of the IGBT is determined by the collector-emitter saturation voltage drop. V ce collector current i c,I It is calculated from the duty cycle D, and the formula is:

[0017] In the formula, T I For the junction temperature; V ce0 This represents the initial saturation pressure drop; r ce For on-resistance; V ce,25℃ This represents the saturation voltage drop of the IGBT at a temperature of 25°C. r ce , 25℃ The on-resistance of the IGBT at 25°C; TC V and TC r They represent V ce and r ce Temperature coefficient; D (t) is the duty cycle function (time domain); kpi The value is 0.965; k di The value is 0.00065; k pv The value is 0.003; T i The value is 0.0045; △ U It is the difference between the input DC voltage and the reference voltage; Let be the impulse function; It is a unit step function; Because there is an overlap between voltage and current during the IGBT's turn-on and turn-off states, switching losses occur. The calculation formula is as follows:

[0018] In the formula, f sw The switching frequency; E on,I , E off,I These represent the energy losses during IGBT turn-on and turn-off, respectively. U dc , i c,I The operating voltage and current of the IGBT; U N , i N Rated voltage and current; K on,I , K off,I This is the coefficient representing the influence of the IGBT gate resistance on turn-on and turn-off losses. K T,I This is the temperature correction factor for switching energy loss. IGBT total power loss P I for:

[0019] FWD's flux loss P CL,F It is due to its saturation pressure drop V F and on-state current i c,F Calculations show that, since the FWD and IGBT alternately conduct, the conduction loss of the FWD is... P CL,F for:

[0020] In the formula, TF is the FWD junction temperature; The turn-off loss of the FWD is caused by the reverse recovery current of the diode, and its calculation formula is as follows:

[0021] In the formula, E off,F The energy loss due to the reverse recovery of FWD; K off,F The coefficient representing the influence of the forward resistance of the FWD on the turn-off loss; K T,F The temperature coefficient for shutting off energy loss; Total loss of FWD P F for:

[0022] The total loss of the DC / DC converter is:

[0023] In the formula, N BFBIC This refers to the number of BFBIC submodules in the DC / DC converter. Transmission line power loss:

[0024] In the formula, I out This refers to the output current of a single fan. R c The resistance of each line segment; r c Resistance per unit length of the line; l c The length of each line segment; The power loss of MMC is: The on-state loss PCL,I,X,M of an IGBT is calculated from the collector-emitter saturation voltage drop Vce,X,M and the collector current ic,I,X,M in the Xth operating state (on / off) during the Mth operating phase of a single power frequency cycle. The formula is as follows:

[0025] In the formula, T I,X,M For the first M The first of the work phases X Junction temperature during operation; V ce0,X,M For the first M The first of the work phases X Initial saturation pressure drop during operation; r ce,X,M For the firstM The first of the work phases X On-resistance during operation; V ce,X,M,25℃ For the first M The first of the work phases X The saturation voltage drop of the IGBT at 25°C during operation; r ce,X,M,25℃ For the first M The first of the work phases X The on-resistance of the IGBT at 25°C during operation; TC V,X,M and TC r,X,M They represent the first M The first of the work phases X within the state V ce,X,M and r ce,X,M Temperature coefficient; The formulas for calculating IGBT switching losses under different operating conditions are as follows:

[0026] In the formula, E on,I,X,M , E off,I,X,M The first M The first of the work phases X Energy loss during IGBT turn-on and turn-off in operation; m on and m off They represent time respectively T The number of times the internal IGBT is switched on; Total power loss of IGBT under different operating conditions P I,X,M for:

[0027] Diode conduction losses under different operating conditions P CL,F,X,M It is due to its saturation pressure drop V F,X,M and on-state current i c,F,X,M The calculated on-state loss P CL,F,X,M for:

[0028] In the formula, T F,X,M For the first M The first of the work phasesX Diode junction temperature during operation; The formulas for calculating the turn-off loss of a diode under different operating conditions are as follows:

[0029] In the formula, E off,F,X,M For the first M The first of the work phases X Energy loss during reverse recovery of the diode in operating state; n off These represent the number of times the diode switches within time T; Total losses of diodes under different operating conditions P F,X,M for:

[0030] The total loss of MMC is:

[0031] In the formula, N SM This represents the number of bridge arm submodules in the MMC converter. The power loss of the transformer is:

[0032] In the formula, P 0 represents no-load loss in kW; P K Rated load loss (kW); K T This is the load fluctuation loss coefficient; β This represents the average load factor of the transformer.

[0033] Furthermore, the reliability index of the DC wind farm is calculated; Probability of power supply shortage PPSD

[0034] In the formula, S The set of system states that cannot provide power support within a given time interval; T The number of hours within a given time interval; t sta System status sta The duration for which power support cannot be provided; Power supply outage time DPSD

[0035] Power supply shortage APSD

[0036] In the formula, C sta The system is in a state sta Power shortage; F sta The system is in a state sta Lower power loss frequency; t sta System status sta The duration for which power support cannot be provided; Electricity supply APS

[0037] In the formula, E sta To account for the impact of losses, the system is in a certain state. sta Power supply capacity; P LOSS,X In order to be in X DC loss in DC wind farm topology; P LOSS,SP DC loss in a series-parallel DC wind farm topology; P LOSS,P DC loss in a parallel DC wind farm topology; P wt,loss For fan losses; P DC / DC,loss DC / DC losses; P MMC,loss For MMC loss; P cable,loss For transmission line loss; P CDC / DC,loss For centralized DC / DC losses; The power collection system supplies the missing power. APSDCS

[0038] In the formula, C staZ,JD When the wind turbine is at full power, the collector system is in a certain state. staZ JD power shortage power; F staZ,JD The system is in a state staZ ,JD power loss frequency; t staZ,JD System status staZ JD cannot provide power for the duration of the power supply. Send out the system to supply the missing power APSDTS

[0039] In the formula, C staZ,SC When the wind turbine is at full power, the power transmission system is in a state of readiness. staZ Power shortage under SC; F staZ,SC The system is in a state staZ , SC power loss frequency; t staZ,SC System status staZ The duration for which SC cannot provide power support; Wind turbine failure rate WTFR

[0040] In the formula, NU i,WF The number of wind turbine failures in each state; NU z,WF This represents the total number of wind turbine units. NU s,WF Number of fault states of wind turbine units; Equipment failure rate EFR

[0041] In the formula, NU i,EQ The number of equipment failures (excluding fans) within each state; NU z,EQ This refers to the total number of equipment (excluding fans); NU s,EQ This represents the number of equipment (excluding fans) in fault condition.

[0042] Furthermore, the normal cloud generator is used to generate cloud droplets corresponding to the cloud models for each reliability index; based on the distance between the cloud droplets and the desired outcome and the membership degree of the cloud droplets, the average influence value of the cloud droplets corresponding to each index for the cloud model is calculated, and the average influence value of the cloud droplets for each cloud model is normalized to obtain the weights of the corresponding reliability indices, including: Reliability index weight calculation based on cloud droplet impact value integrated cloud model; The cloud model uses three data points to represent its characteristics: (1) Expectation: The expectation of the distribution of cloud droplets in the domain space is generally represented by the symbol Ex; (2) Entropy: The degree of uncertainty, which is determined by the degree of outlier and the degree of ambiguity, and is represented by the symbol En; (3) Hyperentropy: used to measure the uncertainty of entropy, that is, the entropy of entropy, represented by the symbol He; there are two types of cloud generators: forward cloud generator and reverse cloud generator, which are used to generate enough cloud droplets and calculate the cloud digital features (Ex, En, He) of cloud droplets, respectively.

[0043] The triggering mechanism of the positive cloud generator is as follows: Generate with En as the expectation and He as the target. 2 En' is a normally distributed random number with variance. Generate normally distributed random numbers with expectation Ex and variance En'. x ; Calculate membership degree;

[0044] ( x , μ ( x )) is relative to the domain of discourse U A cloud droplet; Repeat the above steps until enough cloud droplets are generated; Based on the above, the formula for calculating the average impact value of cloud droplets is as follows:

[0045] The final weight value is calculated using the following formula:

[0046] In the formula, A i Indicates the first i The average impact value of cloud droplets in a cloud model; M represents the number of cloud models; Establish a comprehensive reliability evaluation index for DC wind farms;

[0047] In the formula, R RCE Comprehensive reliability assessment index; P E This is the system's rated power.

[0048] Secondly, the present invention provides a comprehensive reliability evaluation system for DC wind farms, comprising: The data acquisition module is used to acquire the topology of the DC wind farm, component reliability data, and historical wind speed data; establish a DC wind turbine output status model based on historical wind speed data, and establish a component outage probability model based on component reliability data; randomly sample the output status of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output status model, and calculate the total output power of the DC wind farm. The fault tree model building module is used to build a fault tree model of the DC wind farm by combining the topology of the DC wind farm; based on the fault tree model, it makes logical judgments on the state of each system, analyzes whether the turbine needs to be disconnected and the number of turbines to be disconnected, and counts the total output power of the remaining normally operating turbines in the DC wind farm after the turbines are disconnected. The index calculation module is used to define reliability evaluation indicators from spatial and object perspectives based on the total output power after turbine disconnection and the DC loss of the DC wind farm, and to calculate and form a multi-level reliability index system for the DC wind farm. The weight calculation module is used to generate cloud droplets for cloud models corresponding to each reliability index using a normal cloud generator; based on the distance between the cloud droplet and the expectation and the membership degree of the cloud droplet, it calculates the average influence value of the cloud droplet for each index corresponding to the cloud model, and normalizes the average influence value of the cloud droplet for each cloud model to obtain the weight of the corresponding reliability index. The output module is used to calculate the comprehensive reliability evaluation index of the DC wind farm based on each reliability index and its weight, and to serve as the comprehensive reliability evaluation result of the DC wind farm.

[0049] Thirdly, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the comprehensive reliability assessment method for a DC wind farm.

[0050] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the comprehensive reliability assessment method for a DC wind farm.

[0051] Compared with the prior art, the present invention has the following technical effects: This invention acquires DC wind farm topology, component reliability data, and historical wind speed data, and establishes a DC wind turbine output state model, which can accurately describe the impact of wind speed uncertainty on wind turbine output and provide a basis for subsequent reliability assessment. By establishing a component outage probability model and randomly sampling the output status of DC wind turbines, the random failure process of each component in a wind farm can be simulated, thus improving the realism of the assessment. By combining the topology to establish a fault tree model and making logical judgments on the state of each system, it is possible to accurately analyze the demand for and number of turbines to be cut off, and effectively reflect the actual power supply capacity of the DC wind farm under fault conditions. By defining reliability evaluation indicators based on the total output power and DC loss after the generator is switched off, a multi-level reliability indicator system can be formed from spatial and object aspects, covering the generator end, collection link, equipment end, transmission link and receiving end. By using a normal cloud generator to generate cloud droplets and calculate the average impact value of cloud droplets, we can fully explore the nature of cloud models in handling randomness and fuzziness, and improve the rationality and accuracy of weight calculation. By calculating the comprehensive reliability assessment index, a multi-dimensional comprehensive assessment of the reliability of DC wind farms can be achieved, providing key technical support for the planning, operation and maintenance of DC wind farms and the grid connection of new energy sources. Attached Figure Description

[0052] Figure 1 This is a flowchart of the present invention.

[0053] Figure 2 This is a flowchart of the DC wind farm reliability assessment based on the fault tree-sequential Monte Carlo method of the present invention.

[0054] Figure 3 This is a diagram illustrating the reliability index system of the DC wind farm of this invention.

[0055] Figure 4 This is a graph showing the weight calculation results of the various reliability indicators of this invention.

[0056] Figure 5 This is a diagram showing the comprehensive reliability evaluation results of a DC wind farm according to the present invention. Detailed Implementation

[0057] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0058] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0059] The present invention will now be described in further detail with reference to the accompanying drawings: Example 1, see Figure 1 In one embodiment of the present invention, a comprehensive reliability assessment method for DC wind farms is provided to address the shortcomings of existing comprehensive reliability assessment methods in terms of unclear logic in the construction of the indicator system and lack of rationality in weight calculation.

[0060] Specifically, the comprehensive reliability assessment method for DC wind farms of the present invention includes the following steps: S1: Obtain the topology, component reliability data, and historical wind speed data of the DC wind farm; establish a DC wind turbine output status model based on the historical wind speed data, and establish a component outage probability model based on the component reliability data; randomly sample the output status of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output status model, and calculate the total output power of the DC wind farm.

[0061] The topology of a DC wind farm includes the connections between components such as wind turbines, DC / DC converters, MMC converters, DC cables, and centralized DC / DC converters. Component reliability data includes parameters such as failure rate, repair rate, and mean time between failures (MTBF) for each component. Historical wind speed data is a time series of historical wind speeds in the wind farm's location, used to simulate random fluctuations in wind speed. By establishing a DC wind turbine output state model, the uncertainty of wind speed can be transformed into the uncertainty of wind turbine output, thereby simulating the output fluctuations of the DC wind farm. The component outage probability model describes the probability of each component failing and shutting down under different operating conditions; its construction relies on component reliability data. Random sampling of the DC wind turbine output state is used to statistically simulate the actual operating state of the wind farm, thereby calculating the total output power of the DC wind farm.

[0062] S2: Establish a fault tree model for the DC wind farm based on the topology of the DC wind farm; make logical judgments on the state of each system based on the fault tree model, analyze whether it is necessary to disconnect the turbines and the number of turbines to be disconnected, and count the total output power of the remaining normally operating turbines in the DC wind farm after disconnection.

[0063] A fault tree model is a logical cause-and-effect diagram used to describe how the failure of individual components in a system leads to system failure. By establishing a fault tree model in conjunction with the topology of a DC wind farm, component failures can be linked to system failures. Logical judgments are made for each system state to determine whether and how many turbines need to be tripped under a given system state, thus simulating real-world operational control logic. Statistical analysis of the total output power of the remaining operating turbines after tripping reflects the power that the wind farm can still supply to the grid after a fault occurs.

[0064] S3: Based on the total output power after turbine disconnection and the DC loss of the DC wind farm, define reliability evaluation indicators from the perspectives of space and object, calculate and form a multi-level reliability indicator system for the DC wind farm.

[0065] DC losses in a DC wind farm include power losses from DC wind turbines, DC / DC converters, transmission lines, MMC (Multi-Mechanical Control Unit), centralized DC / DC converters, and transformers. Defining reliability evaluation indicators from spatial and object-specific perspectives aims to cover different dimensions: the generation end, the DC collection system, the equipment end, the transmission end, and the receiving end. Multi-level reliability indicators include the probability of power supply shortage, the duration of power supply shortage, the amount of power supply shortage, the amount of power supplied, the amount of power supply shortage in the collection system, the amount of power supply shortage in the transmission system, the wind turbine failure rate, and the equipment failure rate. These indicators reflect the reliability level of a DC wind farm from different perspectives.

[0066] S4: Use a normal cloud generator to generate cloud droplets for each reliability index corresponding to the cloud model; calculate the average influence value of cloud droplets for each index corresponding to the cloud model based on the distance between the cloud droplets and the expected value and the membership degree of the cloud droplets, and normalize the average influence value of cloud droplets for each cloud model to obtain the weight of the corresponding reliability index.

[0067] The cloud model is a model that describes the concept of uncertainty using three numerical features: expectation, entropy, and hyperentropy. Expectation reflects the cloud droplet center, entropy reflects the degree of uncertainty, and hyperentropy reflects the uncertainty of entropy. The normal cloud generator can generate cloud droplets based on the numerical features of the cloud model, simulating the randomness and fuzziness of the indicator data. The average cloud droplet impact value considers the distance and membership degree between the cloud droplet and the expectation, reflecting the local influence of the cloud droplet on the overall cloud structure. Normalizing the average cloud droplet impact values ​​of each cloud model yields the weights of each reliability indicator, reasonably reflecting the importance of different indicators in the comprehensive evaluation.

[0068] S5: Calculate the comprehensive reliability evaluation index of DC wind farm based on each reliability index and its weight, and use it as the comprehensive reliability evaluation result of DC wind farm.

[0069] The comprehensive reliability assessment index for DC wind farms is obtained by weighting various reliability indicators, and can comprehensively reflect the multi-dimensional reliability level of DC wind farms. This index can be used for planning, operation and maintenance decisions of DC wind farms and for providing technical support for new energy grid connection.

[0070] This invention presents a comprehensive reliability assessment method for DC wind farms. By establishing a fault tree-sequential Monte Carlo model, defining multi-level reliability indices, and weighting the cloud model based on cloud droplet influence values, it achieves a comprehensive assessment of the multi-dimensional reliability of DC wind farms. The steps are sequentially linked: first, foundation and wind speed data are acquired and an output model is established; then, a fault tree is built and state logic judgments are performed; next, reliability indices are calculated; then, the cloud model is used to determine the weights of the indices; and finally, a comprehensive assessment index is output. This process design fully considers the characteristics of complex topology, multiple indices, and high uncertainty in DC wind farms. By weighting based on cloud droplet influence values, it overcomes the limitations of traditional cloud models that rely solely on three numerical features to determine weights, thus improving the rationality and accuracy of the assessment results.

[0071] Example 2: The present invention provides a comprehensive reliability assessment method for DC wind farms, comprising the following steps: A reliability assessment model for DC wind farms is constructed based on the fault tree-sequential Monte Carlo method. The reliability evaluation indicators are defined in terms of space, object, time limit, and degree, forming a multi-level and multi-stage reliability indicator system for DC wind farms. An integrated cloud model method based on cloud droplet impact value is introduced, and the randomness and fuzziness of index data are processed by a normal cloud generator to achieve a comprehensive assessment of the multi-dimensional reliability of DC wind farms.

[0072] The steps for constructing a DC wind farm reliability assessment model based on the fault tree-sequential Monte Carlo method are as follows: Step 1: Wind Turbine Output Model and Cluster Analysis The random variation in wind speed leads to uncertainty in the output power of wind turbines within a wind farm. This paper uses the Weibull distribution to construct a wind speed model. The specific probability density function and cumulative distribution function of this distribution are expressed as follows:

[0073] In the formula, k w The shape parameter of the Weibull distribution can characterize the range of wind speed variation. k w The larger the value, the smaller the distribution range of wind speed; c w is the scale parameter of the Weibull distribution; F A random number that follows a uniform distribution in the range [0,1].

[0074] The conversion relationship between fan output power and wind speed is as follows:

[0075] In the formula, v ciTo cut in wind speed; v co To cut off the wind speed; v R Rated wind speed; P R This refers to the rated power of a single fan.

[0076] As the wind turbine output model shows, the uncertainty of wind speed leads to the randomness of the turbine output state, making it difficult to accurately predict the wind farm output. Therefore, it is necessary to perform probabilistic statistics on its output state. However, calculating all time-series output states and their corresponding probabilities within a certain time period would involve a massive amount of computation, undoubtedly reducing computational efficiency. Furthermore, different output states often exhibit high similarity; therefore, it is not necessary to statistically analyze every single independent state. Therefore, this paper employs the K-means clustering method to classify and analyze wind power data within a certain time period, aiming to reduce computational complexity and improve analytical efficiency while providing better data support for subsequent reliability assessments of DC wind farms.

[0077] K-means iteratively updates cluster centers and redistributes data points, ultimately dividing the dataset into K clusters. Data points within each cluster are relatively close in the feature space, while data points in different clusters are relatively far apart. The specific process is as follows: (1) Input all wind turbine output datasets Assuming the maximum number of clusters is Kmax, and the initial number of clusters is... k ( k ≤Kmax), that is, dividing the output data into ,common k Clusters, and randomly selected k Initial point As the initial cluster center.

[0078] (2) Calculate the output data With each cluster center Manhattan distance between them:

[0079] (3) Calculate p i and k After the distance of each cluster center, p i Assign it to the cluster corresponding to the nearest cluster center; (4) Recalculate the cluster centers based on the output data contained in each cluster after the update:

[0080] In the formula, cj For the updated cluster C j The new cluster center; For the updated cluster C j The number of output data.

[0081] (5) Repeat steps (2) to (4) until the change in the cluster center value of each cluster is less than the set threshold or the set maximum number of iterations is reached.

[0082] (6) Calculate the average silhouette coefficient, which ranges from [-1, 1]. The closer it is to 1, the better the clustering effect. Its expression is:

[0083] In the formula, SC k When the data is divided into k The average profile coefficient for a single cluster, a i for p i The average distance to other data in the same cluster. b i for p i The average distance to all data in the nearest cluster.

[0084] (7) Compare the differences k The value of SC, when the maximum value of SC is selected. k The value is the optimal number of clusters K, which is used to count the number of output data and the cluster center value in each cluster.

[0085] (8) Output the cluster centers of K clusters. c j and state probability ,in Divide the number of output data points in each cluster by the total number of output data points, i.e.:

[0086] Step 2: Construct a DC wind farm reliability assessment model based on the fault tree-sequential Monte Carlo method. (1) Input data related to the topology and component reliability of the DC wind farm; (2) Establish a DC wind turbine output state model based on historical wind speed data, and establish a component shutdown probability model in combination with equipment reliability parameters; (3) Randomly sample the output status of DC wind turbines in the system, combine the output status model results of DC wind turbines in (2), determine the output of each wind turbine, and calculate the total output power of the DC wind farm in this sampling. (4) Establish a fault tree model of the DC wind farm based on the topology of the DC wind farm; (5) The sequential Monte Carlo simulation method is used to sample the working state of each component in the DC wind farm to obtain the time-series state transition sequence of each component and the overall time-series state transition sequence of the system. (6) Based on the aforementioned fault tree model of DC wind farm, the system state is logically judged in sequence, and it is analyzed whether the system needs to disconnect the turbine and the number of turbines to be disconnected under the state. (7) Calculate the total output power of the remaining normally operating wind turbines in the DC wind farm after the turbines are switched off; (8) Calculate the reliability index of DC wind farm; (9) Determine whether the reliability index meets the set calculation period. If it does, output the reliability index of the DC wind farm; otherwise, return to step (2). The evaluation process is as follows: Figure 2 As shown.

[0087] The steps to establish a multi-level, multi-stage reliability index system for DC wind farms, defining hierarchical, object-based, time-limited, and degree-based reliability evaluation indicators from spatial and object perspectives, are as follows: Step 1: Analysis and calculation of DC losses in DC wind farms; (1) The power loss of a DC fan is:

[0088] In the formula, P g,loss For generator losses; P sw,vsc For converter switching losses; P c,vsc This refers to the converter's conduction loss. f s This refers to the converter switching frequency; E von,I , E on,I These are the IGBT turn-on losses and turn-off losses within the converter, respectively. E off,D This refers to the diode turn-off loss within the converter; U v,ref This is the rated voltage of the converter; I v,ref This is the rated current of the converter; I wt This represents the output current of a single fan. U wt This refers to the output voltage of a single fan. N vsc This refers to the number of IGBTs in the converter; P vcond,I ,P vcond,D These represent the conduction losses of the IGBTs and diodes within the converter, respectively.

[0089] (2) The power loss of the DC / DC converter is: IGBT conduction loss P CL,I It is caused by the saturation voltage drop of the emitter V ce collector current i c,I It is calculated from the duty cycle D, and the formula is:

[0090] In the formula, T I For the junction temperature; V ce0 This represents the initial saturation pressure drop; r ce For on-resistance; V ce,25℃ This represents the saturation voltage drop of the IGBT at a temperature of 25°C. r ce , 25℃ The on-resistance of the IGBT at 25°C; TC V and TC r They represent V ce and r ce Temperature coefficient; D (t) is the duty cycle function (time domain); k pi The value is 0.965; k di The value is 0.00065; k pv The value is 0.003; T i The value is 0.0045; △ U It is the difference between the input DC voltage and the reference voltage; Let be the impulse function; It is a unit step function.

[0091] Because there is an overlap between voltage and current during the IGBT's turn-on and turn-off states, switching losses occur. The calculation formula is as follows:

[0092] In the formula, f sw The switching frequency;E on,I , E off,I These represent the energy losses during IGBT turn-on and turn-off, respectively. U dc , i c,I The operating voltage and current of the IGBT; U N , i N Rated voltage and current; K on,I , K off,I This is the coefficient representing the influence of the IGBT gate resistance on turn-on and turn-off losses. K T,I This is the temperature correction factor for switching energy loss.

[0093] IGBT total power loss P I for:

[0094] FWD's flux loss P CL,F It is due to its saturation pressure drop V F and on-state current i c,F Calculations show that, since the FWD and IGBT alternately conduct, the conduction loss of the FWD is... P CL,F for:

[0095] In the formula, T F This is the junction temperature of the FWD (Foil-Dwelling) junction.

[0096] The turn-off loss of the FWD is caused by the reverse recovery current of the diode, and its calculation formula is as follows:

[0097] In the formula, E off,F The energy loss due to the reverse recovery of FWD; K off,F The coefficient representing the influence of the forward resistance of the FWD on the turn-off loss; K T,F The temperature coefficient for shutting off energy loss.

[0098] Total loss of FWD P F for:

[0099] The total loss of the DC / DC converter is:

[0100] In the formula, N BFBIC This represents the number of BFBIC submodules in the DC / DC converter.

[0101] (3) Transmission line power loss:

[0102] In the formula, I out This refers to the output current of a single fan. R c The resistance of each line segment; r c Resistance per unit length of the line; l c This represents the length of each line segment.

[0103] (4) The power loss of MMC is: IGBT conduction loss P CL,I,X,M It is the first time under a single power frequency cycle M The first of the work phases X Emitter saturation voltage drop in each working state (engaged / disengaged) V ce,X,M collector current i c,I,X,M The formula is calculated as follows:

[0104] In the formula, T I,X,M For the first M The first of the work phases X Junction temperature during operation; V ce0,X,M For the first M The first of the work phases X Initial saturation pressure drop during operation; r ce,X,M For the first M The first of the work phases X On-resistance during operation; V ce,X,M,25℃ For the first M The first of the work phases X The saturation voltage drop of the IGBT at 25°C during operation; r ce,X,M,25℃ For the first M The first of the work phases XThe on-resistance of the IGBT at 25°C during operation; TC V,X,M and TC r,X,M They represent the first M The first of the work phases X within the state V ce,X,M and r ce,X,M Temperature coefficient.

[0105] The formulas for calculating IGBT switching losses under different operating conditions are as follows:

[0106] In the formula, E on,I,X,M , E off,I,X,M The first M The first of the work phases X Energy loss during IGBT turn-on and turn-off in operation; m on and m off They represent time respectively T The number of times the internal IGBT is switched on and off.

[0107] Total power loss of IGBT under different operating conditions P I,X,M for:

[0108] Diode conduction losses under different operating conditions P CL,F,X,M It is due to its saturation pressure drop V F,X,M and on-state current i c,F,X,M The calculated on-state loss P CL,F,X,M for:

[0109] In the formula, T F,X,M For the first M The first of the work phases X Diode junction temperature during operation.

[0110] The formulas for calculating the turn-off loss of a diode under different operating conditions are as follows:

[0111] In the formula, E off,F,X,M For the firstM The first of the work phases X Energy loss during reverse recovery of the diode in operating state; n off These represent the number of times the diode switches within time T.

[0112] Total losses of diodes under different operating conditions P F,X,M for:

[0113] The total loss of MMC is:

[0114] In the formula, N SM This represents the number of bridge arm submodules in the MMC converter.

[0115] (5) The power loss of the centralized DC / DC converter is: The centralized DC / DC converter topology constructed in this paper consists of two modular multilevel converters connected by an isolation transformer. In the design of the MMC converter arms at both the input and output ends, submodules are connected in series to effectively alleviate the requirements for drive consistency and voltage equalization faced by directly connecting a large number of power devices in high-voltage applications. The power loss generated by the MMC converters at both the input and output ends during operation mainly originates from the power devices. This loss includes conduction losses and switching losses. The loss assessment method is the same as the calculation steps for the power loss of the aforementioned MMC converter, and will not be repeated here.

[0116] The power loss of the transformer is:

[0117] In the formula, P 0 represents no-load loss (kW); P K Rated load loss (kW); K T This is the load fluctuation loss coefficient; β This represents the average load factor of the transformer.

[0118] Step 2: Calculation of reliability indicators for DC wind farms; (1) Probability of Power Supply Deficiency PPSD )

[0119] In the formula, SThe set of system states that cannot provide power support within a given time interval; T The number of hours within a given time interval; t sta System status sta Duration during which power cannot be provided.

[0120] (2) Duration of Power Supply Deficiency DPSD )

[0121] (3) Amount of Power Supply Deficiency APSD )

[0122] In the formula, C sta The system is in a state sta Power shortage; F sta The system is in a state sta Lower power loss frequency; t sta System status sta Duration during which power cannot be provided.

[0123] (4) Amount of Power Supply APS )

[0124] In the formula, E sta To account for the impact of losses, the system is in a certain state. sta Power supply capacity; P LOSS,X In order to be in X DC loss in DC wind farm topology; P LOSS,SP DC loss in a series-parallel DC wind farm topology; P LOSS,P DC loss in a parallel DC wind farm topology; P wt,loss For fan losses; P DC / DC,loss DC / DC losses; P MMC,loss For MMC loss; P cable,loss For transmission line loss;P CDC / DC,loss This refers to the loss of a centralized DC / DC converter.

[0125] (5) Amount of Power Supply Deficiency of the Collector System APSDCS )

[0126] In the formula, C staZ,JD When the wind turbine is at full power, the collector system is in a certain state. staZ JD power shortage power; F staZ,JD The system is in a state staZ ,JD power loss frequency; t staZ,JD System status staZ JD cannot provide power for the duration of the power supply.

[0127] (6) Amount of Power Supply Deficiency of the Transmission-out System APSDTS )

[0128] In the formula, C staZ,SC When the wind turbine is at full power, the power transmission system is in a state of readiness. staZ Power shortage under SC; F staZ,SC The system is in a state staZ , SC power loss frequency; t staZ,SC System status staZ SC cannot provide power support for the duration of time.

[0129] (7) Wind Turbine Failure Ratio WTFR )

[0130] In the formula, NU i,WF The number of wind turbine failures in each state; NU z,WF This represents the total number of wind turbine units. NU s,WF This represents the number of fault states of the wind turbine unit.

[0131] (8) Equipment Failure Ratio EFR )

[0132] In the formula, NU i,EQ The number of equipment failures (excluding fans) within each state; NU z,EQ This refers to the total number of equipment (excluding fans); NU s,EQ This represents the number of equipment (excluding fans) in fault condition.

[0133] The steps for comprehensively evaluating the multi-dimensional reliability of DC wind farms by introducing an integrated cloud model method based on cloud droplet impact values ​​and processing the randomness and fuzziness of index data through a normal cloud generator are as follows: Step 1: Establish a comprehensive reliability evaluation index system for DC wind farms; The wind turbine failure ratio focuses on the power supply layer, providing a direct quantification of the operational stability of the wind farm's generation side and reflecting the overall health of the wind turbine cluster. The equipment failure ratio, from the perspective of collaboration between the equipment and power receiving layers, covers failures in core equipment such as DC transmission lines and converter stations, effectively demonstrating the comprehensive impact of equipment operation on system reliability. From the power receiving layer's perspective, power supply failure probability (PPSD), power supply failure duration (DPSD), power supply failure amount (APSD), and power supply amount (APS) are established to quantitatively characterize the system's power supply reliability. PPSD reflects the probability that the DC wind farm cannot meet the load demand of the power receiving layer during operation, quantifying the frequency of power shortage events. DPSD precisely measures the duration of each power shortage event, providing a time-based basis for assessing the impact of power outages on users. APSD further clarifies the actual amount of power shortage during power shortage events, intuitively showing the scale of energy loss caused by power outages. APS characterizes the actual amount of power delivered by the system to the power receiving layer after considering factors such as DC losses, reflecting the system's power supply capacity under normal operating conditions. In the power collection stage, APSDJD is used to quantify the risk of coordinated failure between wind turbines and DC power collection lines; in the power transmission stage, APSDSC is used to assess the impact of converter station and DC line failures on power transmission capacity. This two-dimensional analysis model can deconstruct the reliability problem of complex systems into quantifiable subsystem risks, and through risk matrix and sensitivity analysis, clearly present the contribution of failures in each stage to the overall power supply reliability.

[0134] This assessment system, through a logical chain of failure occurrence mechanism, impact process, and final result, forms a closed-loop analysis path of "risk identification - impact quantification - capability assessment," providing comprehensive quantitative support for system reliability analysis, optimization design, and operational decisions. It provides important basis for improving and optimizing the reliability of DC wind farms. The indicator system is as follows: Figure 3 As shown.

[0135] Step 2: Calculate the reliability index weights of the cloud model based on cloud droplet impact values; The cloud model uses three data points to represent its characteristics: (1) Expectation: the expectation of the distribution of cloud droplets in the domain space, generally represented by the symbol Ex. (2) Entropy: the degree of uncertainty, determined by the degree of outlier and the degree of fuzziness, generally represented by the symbol En. (3) Hyperentropy: used to measure the uncertainty of entropy, i.e., the entropy of entropy, generally represented by the symbol He. There are two types of cloud generators: forward cloud generator and reverse cloud generator, which are used to generate enough cloud droplets and calculate the cloud digital features (Ex, En, He) of the cloud droplets, respectively.

[0136]

[0137] The triggering mechanism of the positive cloud generator is as follows: (1) Generate a value with En as the desired value and He as the desired value. 2 En' is a normally distributed random number with variance. (2) Generate normally distributed random numbers with expectation Ex and variance En'. x ; (3) Calculate the membership degree;

[0138] ( x , μ ( x )) is relative to the domain of discourse U A cloud droplet.

[0139] (4) Repeat the above steps until enough cloud droplets are generated.

[0140] (5) Based on the above, the formula for calculating the average impact value of cloud droplets is as follows:

[0141] (6) The final weight value is calculated using the following formula:

[0142] In the formula, A i Indicates the first i The average impact value of cloud droplets in each cloud model; M represents the number of cloud models. The weight calculation results of each reliability index are as follows: Figure 4 As shown.

[0143] Step 3: Establish comprehensive reliability evaluation indicators for DC wind farms;

[0144] In the formula, R RCE Comprehensive reliability assessment index; P E This is the system's rated power.

[0145] Substituting the calculated weight values ​​into the comprehensive reliability evaluation index for DC wind farms, the final comprehensive evaluation result can be obtained, such as... Figure 5 As shown.

[0146] In another embodiment of the present invention, a comprehensive reliability assessment system for DC wind farms is provided, which can be used to implement the above-mentioned comprehensive reliability assessment method for DC wind farms. Specifically, the comprehensive reliability assessment system for DC wind farms includes: The data acquisition module is used to acquire the topology of the DC wind farm, component reliability data, and historical wind speed data; establish a DC wind turbine output status model based on historical wind speed data, and establish a component outage probability model based on component reliability data; randomly sample the output status of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output status model, and calculate the total output power of the DC wind farm. The fault tree model building module is used to build a fault tree model of the DC wind farm by combining the topology of the DC wind farm; based on the fault tree model, it makes logical judgments on the state of each system, analyzes whether the turbine needs to be disconnected and the number of turbines to be disconnected, and counts the total output power of the remaining normally operating turbines in the DC wind farm after the turbines are disconnected. The index calculation module is used to define reliability evaluation indicators from spatial and object perspectives based on the total output power after turbine disconnection and the DC loss of the DC wind farm, and to calculate and form a multi-level reliability index system for the DC wind farm. The weight calculation module is used to generate cloud droplets for cloud models corresponding to each reliability index using a normal cloud generator; based on the distance between the cloud droplet and the expectation and the membership degree of the cloud droplet, it calculates the average influence value of the cloud droplet for each index corresponding to the cloud model, and normalizes the average influence value of the cloud droplet for each cloud model to obtain the weight of the corresponding reliability index. The output module is used to calculate the comprehensive reliability evaluation index of the DC wind farm based on each reliability index and its weight, and to serve as the comprehensive reliability evaluation result of the DC wind farm.

[0147] The module division in this embodiment of the invention is illustrative and represents only one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the invention can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0148] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions from the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used in the operation of a comprehensive reliability assessment method for DC wind farms.

[0149] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the comprehensive reliability assessment method for DC wind farms in the above embodiments.

[0150] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied 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.

[0151] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0152] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0153] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A comprehensive reliability assessment method for DC wind farms, characterized in that, include: Acquire the topology, component reliability data, and historical wind speed data of the DC wind farm; establish a DC wind turbine output state model based on the historical wind speed data, and establish a component outage probability model based on the component reliability data; randomly sample the output state of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output state model, and calculate the total output power of the DC wind farm. A fault tree model of a DC wind farm is established based on the topology of the DC wind farm. Logical judgments are made on the state of each system based on the fault tree model to analyze whether the wind turbines need to be disconnected and the number of wind turbines to be disconnected. The total output power of the remaining normally operating wind turbines in the DC wind farm after the disconnection is calculated. Based on the total output power after turbine disconnection and the DC loss of the wind farm, reliability evaluation indicators are defined from the aspects of space and object, and a multi-level reliability index system for DC wind farms is calculated and formed. The cloud droplets corresponding to each reliability index are generated using a normal cloud generator. Based on the distance between the cloud droplets and the desired outcome and the membership degree of the cloud droplets, the average influence value of the cloud droplets corresponding to each index is calculated, and the average influence value of the cloud droplets of each cloud model is normalized to obtain the weight of the corresponding reliability index. Based on each reliability index and its weight, a comprehensive reliability evaluation index for DC wind farms is calculated, which serves as the comprehensive reliability evaluation result for DC wind farms.

2. The comprehensive reliability assessment method for DC wind farms according to claim 1, characterized in that, The process involves acquiring DC wind farm topology, component reliability data, and historical wind speed data; and establishing a DC wind turbine output state model based on the historical wind speed data, including: The wind speed model is constructed using the Weibull distribution, whose specific probability density function and cumulative distribution function are expressed as follows: In the formula, The shape parameter of the Weibull distribution can characterize the range of wind speed variation. The larger the value, the smaller the distribution range of wind speed; is the scale parameter of the Weibull distribution; F is a random number that follows a uniform distribution in [0,1]. The conversion relationship between fan output power and wind speed is as follows: In the formula, v ci To cut in wind speed; v co To cut off the wind speed; v R Rated wind speed; P R This refers to the rated power of a single fan. K-means clustering was used to classify and analyze wind power data over a certain time period. K-means iteratively updates cluster centers and redistributes data points, ultimately dividing the dataset into K clusters. Data points within each cluster are relatively close in the feature space, while data points in different clusters are relatively far apart. The specific process is as follows: (1) Input all wind turbine output datasets Assuming the maximum number of clusters is Kmax, and the initial number of clusters is k (k≤Kmax), the output data is divided into... There are k clusters in total, and k initial points are randomly selected. As the initial cluster center; (2) Calculate the output data With each cluster center Manhattan distance between them: (3) Calculate After determining the distances to the k cluster centers, Assign it to the cluster corresponding to the nearest cluster center; (4) Recalculate the cluster centers based on the output data contained in each cluster after the update: In the formula, For the updated cluster The new cluster center; For the updated cluster The number of output data; (5) Repeat steps (2) to (4) until the change in the cluster center value of each cluster is less than the set threshold or the set maximum number of iterations is reached; (6) Calculate the average silhouette coefficient, which ranges from [-1, 1]. The closer it is to 1, the better the clustering effect. Its expression is: In the formula, The mean silhouette coefficient is the number of data clusters divided into k clusters. for The average distance to other data in the same cluster. for The average distance to all data in the nearest cluster; (7) Compare the size of SC under different k values, select the k value corresponding to the maximum value of SC as the optimal number of clusters K, and count the number of output data and cluster center value contained in each cluster. (8) Output the cluster centers cj of the K clusters and their state probabilities. ,in Divide the number of output data points in each cluster by the total number of output data points, i.e.: 。 3. The comprehensive reliability assessment method for DC wind farms according to claim 1, characterized in that, The component downtime probability model is established by combining component reliability data; the output status of the DC wind turbines in the system is randomly sampled, and the output of each wind turbine is determined by combining the DC wind turbine output status model, and the total output power of the DC wind farm is calculated, including: Input data related to the topology and component reliability of the DC wind farm; A power output state model for DC wind turbines was established based on historical wind speed data, and a component downtime probability model was established in conjunction with equipment reliability parameters. The output status of the DC wind turbines in the system is randomly sampled. Based on the output status model results of the DC wind turbines, the output of each wind turbine is determined, and the total output power of the DC wind farm in this sampling is calculated.

4. The comprehensive reliability assessment method for DC wind farms according to claim 1, characterized in that, The fault tree model of the DC wind farm is established by combining the topology of the DC wind farm; based on the fault tree model, the state of each system is logically judged to establish the fault tree model of the DC wind farm. The sequential Monte Carlo simulation method was used to sample the operating states of each component in the DC wind farm, and the time-series state transition sequence of each component and the overall time-series state transition sequence of the system were obtained. Based on the aforementioned fault tree model of DC wind farm, the system state is logically judged in sequence, and it is analyzed whether the system needs to disconnect the turbines and the number of turbines to be disconnected under the state. Calculate the total output power of the remaining normally operating wind turbines in the DC wind farm after the turbines are switched off. Calculate the reliability indicators of a DC wind farm; Determine whether the reliability index meets the set calculation period. If it does, output the reliability index of the DC wind farm; otherwise, return.

5. The comprehensive reliability evaluation method for DC wind farms according to claim 1, characterized in that, The calculation based on the total output power after turbine cutoff and the DC loss of the DC wind farm includes: DC fan power loss: In the formula, P g,loss For generator losses; P sw,vsc For converter switching losses; P c,vsc This refers to the converter's conduction loss. f s This refers to the converter switching frequency; E von,I , E on,I These are the IGBT turn-on losses and turn-off losses within the converter, respectively. E off,D This refers to the diode turn-off loss within the converter; U v,ref This is the rated voltage of the converter; I v,ref This is the rated current of the converter; I wt This represents the output current of a single fan. U wt This refers to the output voltage of a single fan. N vsc This refers to the number of IGBTs in the converter; P vcond,I , P vcond,D These are the conduction losses of the IGBTs and diodes within the converter, respectively. DC / DC converter power loss: The on-state loss PCL, I of the IGBT is determined by the collector-emitter saturation voltage drop. V ce collector current i c,I It is calculated from the duty cycle D, and the formula is: In the formula, T I For the junction temperature; V ce0 This represents the initial saturation pressure drop; r ce For on-resistance; V ce,25℃ This represents the saturation voltage drop of the IGBT at a temperature of 25°C. r ce , 25℃ The on-resistance of the IGBT at 25°C; TC V and TC r They represent V ce and r ce Temperature coefficient; D (t) is the duty cycle function (time domain); k pi The value is 0.965; k di The value is 0.00065; k pv The value is 0.003; T i The value is 0.0045; △ U It is the difference between the input DC voltage and the reference voltage; Let be the impulse function; It is a unit step function; Because there is an overlap between voltage and current during the IGBT's turn-on and turn-off states, switching losses occur. The calculation formula is as follows: In the formula, f sw The switching frequency; E on,I , E off,I These represent the energy losses during IGBT turn-on and turn-off, respectively. U dc , i c,I The operating voltage and current of the IGBT; U N , i N Rated voltage and current; K on,I , K off,I This is the coefficient representing the influence of the IGBT gate resistance on turn-on and turn-off losses. K T,I This is the temperature correction factor for switching energy loss. IGBT total power loss P I for: FWD's flux loss P CL,F It is due to its saturation pressure drop V F and on-state current i c,F Calculations show that, since the FWD and IGBT alternately conduct, the conduction loss of the FWD is... P CL,F for: In the formula, TF is the FWD junction temperature; The turn-off loss of the FWD is caused by the reverse recovery current of the diode, and its calculation formula is as follows: In the formula, E off,F The energy loss due to the reverse recovery of FWD; K off,F The coefficient representing the influence of the forward resistance of the FWD on the turn-off loss; K T,F The temperature coefficient for shutting off energy loss; Total loss of FWD P F for: The total loss of the DC / DC converter is: In the formula, N BFBIC This refers to the number of BFBIC submodules in the DC / DC converter. Transmission line power loss: In the formula, I out This refers to the output current of a single fan. R c The resistance of each line segment; r c Resistance per unit length of the line; l c The length of each line segment; The power loss of MMC is: The on-state loss PCL,I,X,M of an IGBT is calculated from the collector-emitter saturation voltage drop Vce,X,M and the collector current ic,I,X,M in the Xth operating state (on / off) during the Mth operating phase of a single power frequency cycle. The formula is as follows: In the formula, T I,X,M For the first M The first of the work phases X Junction temperature during operation; V ce0,X,M For the first M The first of the work phases X Initial saturation pressure drop during operation; r ce,X,M For the first M The first of the work phases X On-resistance during operation; V ce,X,M,25℃ For the first M The first of the work phases X The saturation voltage drop of the IGBT at 25°C during operation; r ce,X,M,25℃ For the first M The first of the work phases X The on-resistance of the IGBT at 25°C during operation; TC V,X,M and TC r,X,M They represent the first M The first of the work phases X within the state V ce,X,M and r ce,X,M Temperature coefficient; The formulas for calculating IGBT switching losses under different operating conditions are as follows: In the formula, E on,I,X,M , E off,I,X,M The first M The first of the work phases X Energy loss during IGBT turn-on and turn-off in operation; m on and m off They represent time respectively T The number of times the internal IGBT is switched on; Total power loss of IGBT under different operating conditions P I,X,M for: Diode conduction losses under different operating conditions P CL,F,X,M It is due to its saturation pressure drop V F,X,M and on-state current i c,F,X,M The calculated on-state loss P CL,F,X,M for: In the formula, T F,X,M For the first M The first of the work phases X Diode junction temperature during operation; The formulas for calculating the turn-off loss of a diode under different operating conditions are as follows: In the formula, E off,F,X,M For the first M The first of the work phases X Energy loss during reverse recovery of the diode in operating state; n off These represent the number of times the diode switches within time T; Total losses of diodes under different operating conditions P F,X,M for: The total loss of MMC is: In the formula, N SM This represents the number of bridge arm submodules in the MMC converter. The power loss of the transformer is: In the formula, P 0 represents no-load loss in kW; P K Rated load loss (kW); K T This is the load fluctuation loss coefficient; β This represents the average load factor of the transformer.

6. The comprehensive reliability assessment method for DC wind farms according to claim 1, characterized in that, The calculation of the reliability index of the DC wind farm; Probability of power supply shortage PPSD In the formula, S The set of system states that cannot provide power support within a given time interval; T The number of hours within a given time interval; t sta System status sta The duration for which power support cannot be provided; Power supply outage time DPSD Power supply shortage APSD In the formula, C sta The system is in a state sta Power shortage; F sta The system is in a state sta Lower power loss frequency; t sta System status sta The duration for which power support cannot be provided; Electricity supply APS In the formula, E sta To account for the impact of losses, the system is in a certain state. sta Power supply capacity; P LOSS,X In order to be in X DC loss in DC wind farm topology; P LOSS,SP DC loss in a series-parallel DC wind farm topology; P LOSS,P DC loss in a parallel DC wind farm topology; P wt,loss For fan losses; P DC / DC,loss DC / DC losses; P MMC,loss For MMC loss; P cable,loss For transmission line loss; P CDC / DC,loss For centralized DC / DC losses; The power collection system supplies the missing power. APSDCS In the formula, C staZ,JD When the wind turbine is at full power, the collector system is in a certain state. staZ JD power shortage power; F staZ,JD The system is in a state staZ ,JD power loss frequency; t staZ,JD System status staZ JD cannot provide power for the duration of the power supply. Send out the system to supply the missing power APSDTS In the formula, C staZ,SC When the wind turbine is at full power, the power transmission system is in a state of readiness. staZ Power shortage under SC; F staZ,SC The system is in a state staZ , SC power loss frequency; t staZ,SC System status staZ The duration for which SC cannot provide power support; Wind turbine failure rate WTFR In the formula, NU i,WF The number of wind turbine failures in each state; NU z,WF This represents the total number of wind turbine units. NU s,WF Number of fault states of wind turbine units; Equipment failure rate EFR In the formula, NU i,EQ The number of equipment failures (excluding fans) within each state; NU z,EQ This refers to the total number of equipment (excluding fans); NU s,EQ This represents the number of equipment (excluding fans) in fault condition.

7. The comprehensive reliability assessment method for DC wind farms according to claim 1, characterized in that, The process involves using a normal cloud generator to generate cloud droplets corresponding to each reliability index in the cloud model; calculating the average influence value of the cloud droplets for each index based on the distance between the cloud droplets and the desired outcome and the membership degree of the cloud droplets; and normalizing the average influence value of the cloud droplets for each cloud model to obtain the weights of the corresponding reliability indices, including: Reliability index weight calculation based on cloud droplet impact value integrated cloud model; The cloud model uses three data points to represent its characteristics: (1) Expectation: The expectation of the distribution of cloud droplets in the domain space is generally represented by the symbol Ex; (2) Entropy: The degree of uncertainty, which is determined by the degree of outlier and the degree of ambiguity, and is represented by the symbol En; (3) Hyperentropy: used to measure the uncertainty of entropy, that is, the entropy of entropy, represented by the symbol He; there are two types of cloud generators: forward cloud generator and reverse cloud generator, which are used to generate enough cloud droplets and calculate the cloud digital features (Ex, En, He) of cloud droplets, respectively. The triggering mechanism of the positive cloud generator is as follows: Generate with En as the expectation and He as the target. 2 En' is a normally distributed random number with variance. Generate normally distributed random numbers with expectation Ex and variance En'. x ; Calculate membership degree; ( x , μ ( x )) is relative to the domain of discourse U A cloud droplet; Repeat the above steps until enough cloud droplets are generated; Based on the above, the formula for calculating the average impact value of cloud droplets is as follows: The final weight value is calculated using the following formula: In the formula, A i Indicates the first i The average impact value of cloud droplets in a cloud model; M represents the number of cloud models; Establish a comprehensive reliability evaluation index for DC wind farms; In the formula, R RCE Comprehensive reliability assessment index; P E This is the system's rated power.

8. A comprehensive reliability evaluation system for DC wind farms, characterized in that, include: The data acquisition module is used to acquire the topology of the DC wind farm, component reliability data, and historical wind speed data; establish a DC wind turbine output status model based on historical wind speed data, and establish a component outage probability model based on component reliability data; randomly sample the output status of the DC wind turbines in the system, determine the output of each wind turbine based on the DC wind turbine output status model, and calculate the total output power of the DC wind farm. The fault tree model building module is used to build a fault tree model of the DC wind farm by combining the topology of the DC wind farm; based on the fault tree model, it makes logical judgments on the state of each system, analyzes whether the turbine needs to be disconnected and the number of turbines to be disconnected, and counts the total output power of the remaining normally operating turbines in the DC wind farm after the turbines are disconnected. The index calculation module is used to define reliability evaluation indicators from spatial and object perspectives based on the total output power after turbine disconnection and the DC loss of the DC wind farm, and to calculate and form a multi-level reliability index system for the DC wind farm. The weight calculation module is used to generate cloud droplets for cloud models corresponding to each reliability index using a normal cloud generator; based on the distance between the cloud droplet and the expectation and the membership degree of the cloud droplet, it calculates the average influence value of the cloud droplet for each index corresponding to the cloud model, and normalizes the average influence value of the cloud droplet for each cloud model to obtain the weight of the corresponding reliability index. The output module is used to calculate the comprehensive reliability evaluation index of the DC wind farm based on each reliability index and its weight, and to serve as the comprehensive reliability evaluation result of the DC wind farm.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the comprehensive reliability assessment method for DC wind farms as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the comprehensive reliability assessment method for DC wind farms as described in any one of claims 1 to 7.