A wind turbine generator system reliability state evaluation method and system based on an accident directed graph

Abstract

This invention provides a method and system for assessing the reliability of wind turbine generator sets based on a directed fault graph. Addressing the challenges of complex structures and multiple failure modes in wind turbine generator sets, it employs fault tree analysis to establish a directed fault graph model. To address the difficulty of polymorphism at the root nodes of the directed fault graph, it considers the uncertain logical relationships between components and the system within the wind turbine generator set and uses probability values ​​to measure these relationships in a conditional confidence quality table. Then, it takes all importance rankings into account and performs directed fault graph reasoning according to the same importance ranking, selecting the maximum value to complete the multi-fault state reliability assessment of the wind turbine generator set. This invention effectively assesses the reliability of wind turbine generator sets while avoiding the low efficiency and low accuracy problems of the Monte Carlo method and confidence interval method, respectively.

Description

Technical Field

[0001] This invention relates to the field of wind turbine reliability analysis and fault diagnosis technology, and in particular to a method and system for assessing the reliability status of wind turbines based on a directed accident graph. Background Technology

[0002] As a crucial component of renewable energy conversion, wind turbine generators play a vital role in promoting energy transition and achieving sustainable development. However, the reliability of wind turbine generators has always been a key factor restricting their development. Wind turbine generators are characterized by complex structures and multiple failure modes, and existing technologies mostly rely on fault tree models for reliability assessment. However, fault tree analysis requires events to exist in only two states: normal and faulty, and it cannot represent the uncertainty of the logical relationships between events, making it difficult to meet the requirements for refined reliability assessment of wind turbine generators with multiple failure states. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] Based on this, the present invention provides a method and system for assessing the reliability status of wind turbine generator sets based on a directed fault graph, in order to solve the problem of reliability assessment of multiple fault states caused by the polymorphism and uncertainty of fault logic relationships of wind turbine generator sets.

[0005] (II) Technical Solution

[0006] To achieve the above objectives, this invention provides a method for assessing the reliability status of wind turbine generator sets based on a directed accident graph, comprising:

[0007] S1: Establish a directed graph model of wind turbine generator accidents based on fault tree analysis;

[0008] S101: Study the components and failure modes of wind turbine generator sets, and build a fault tree model of wind turbine generator sets based on the components and failure modes.

[0009] S102: Convert the logic gates in the fault tree model into directed arcs, convert the events in the fault tree model into nodes of the fault directed graph, and generate a conditional confidence quality table based on the meaning of the logic gates.

[0010] S103: Combine the directed arcs, accident directed graph nodes, and conditional confidence quality table to establish an accident directed graph model for wind turbine generator sets;

[0011] S2: Based on the state importance method, perform fault polymorphism reliability assessment on the directed graph model of the wind turbine generator set.

[0012] Specifically, S2 includes:

[0013] S201: Analyze the state of the root node in the directed graph model of the accident, and determine the importance of each state of the fault based on the structure of the directed graph of the accident and the research problem;

[0014] S202: Sort the root node states according to their importance, and based on the constraints of the sum of the states, take the maximum value within the acceptable range to evaluate the reliability of each state of the root node.

[0015] S203: Taking all importance rankings into account, perform directed graph reasoning for wind turbine generator set accidents according to the same importance ranking, and take the maximum value from it to realize intelligent reliability state assessment of multi-state faults of wind turbine generator sets and obtain the confidence interval of each state.

[0016] Optionally, the number of node states in S201 can be three types: normal, half-fault, and fault.

[0017] Optionally, in the directed graph model of the wind turbine generator set accident, the child nodes are connected to the parent node through an "OR gate". That is, when any child node is in a fault state, the parent node is in a fault state; when all child nodes are in a fault-free state, the parent node is in a fault-free state; and when any child node is in an uncertain state and the other child nodes are in a fault-free state, the parent node is also in an uncertain state.

[0018] Optionally, the failure probabilities of leaf nodes in each state in the directed graph model of the accident are:

[0019]

[0020] In the formula, w S (w S w1(w1=0,1,2) represents the state of the leaf node, w2(w2=0,1,2) represents the state of the root node R1, and w2(w2=0,1,2) represents the state of the root node R2. This indicates that leaf node S is in w S The probability of a fault state. This represents the probability of the root node R1 being in state w1. This represents the probability of the root node R2 being in state w2. This indicates that when root node R1 is in state w1 and root node R2 is in state w2, leaf node S is in state w. S The probability of failure in a given state.

[0021] Optionally, the constraints for each state of the root node in the directed graph model of the accident are:

[0022]

[0023] Optionally, the formula for calculating the confidence interval of the system in each state in the accident directed graph model is as follows:

[0024]

[0025] In the formula, N represents the number of root nodes (i.e., components); L represents the number of intermediate nodes (i.e., subsystems); π(Y) l ) represents the intermediate node Y l The set of parent nodes of leaf node S; π(S) represents the set of parent nodes of leaf node S.

[0026] Furthermore, this invention provides a wind turbine generator reliability status assessment system based on a fault-directed graph, comprising:

[0027] At least one processor; and at least one memory communicatively connected to said processor, wherein:

[0028] The memory stores program instructions that can be executed by the processor, and the processor can execute any of the above-described methods for assessing the reliability status of wind turbine generators based on accident directed graphs by calling the program instructions.

[0029] Furthermore, the present invention provides a non-transitory computer-readable storage medium storing computer instructions that cause the computer to execute the wind turbine generator reliability status assessment method based on the accident directed graph described above.

[0030] (III) Beneficial Effects

[0031] As can be seen from the above technical solution, the wind turbine generator reliability status assessment and analysis method and system based on accident directed graph proposed in this invention has the following beneficial effects:

[0032] 1. Based on the characteristics of fault trees in terms of topology, and considering the difficulties of complex structures and multiple fault modes in wind turbine generator sets, a directed graph model of faults is established using fault trees. The logic gates in the fault tree are transformed into directed arcs, events into nodes in the directed graph of faults, and logic gates into conditional confidence quality tables, thus completing the transformation process from fault trees to directed graphs of faults. This realizes the establishment of a directed graph model of faults for wind turbine generator sets, which has the characteristics of simple principle and clear structure.

[0033] 2. Addressing the challenges of numerous components and root node polymorphism in wind turbine generators, the Monte Carlo method employs a large sample size for directed graph reasoning in fault analysis, achieving high accuracy but suffering from efficiency limitations. While the confidence interval method analyzes wind turbine generator fault states using only two directed graph reasoning operations, the analysis involves redundant allocation of perceived states, resulting in accuracy deficiencies. This invention, employing directed graph reasoning in fault analysis, obtains confidence intervals for each wind turbine generator state based on state importance. This effectively assesses wind turbine generator reliability while avoiding the low efficiency and low accuracy issues of the Monte Carlo method and the confidence interval method, respectively.

[0034] 3. The root node of the present invention is divided into two or three states, which is also applicable to situations where the number of root node states is inconsistent. Attached Figure Description

[0035] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0036] Figure 1 This is a basic structural diagram of the doubly-fed wind turbine generator set of the present invention;

[0037] Figure 2 This invention provides an electronic system fault tree;

[0038] Figure 3 This is a flowchart illustrating the process of establishing a directed graph model of accidents based on fault tree analysis in this invention.

[0039] Figure 4 This invention provides a directed graph model of the accident.

[0040] Figure 5 This is a schematic diagram showing the possible ranges of each state when the root node of the present invention has two states;

[0041] Figure 6 This is a schematic diagram showing the possible ranges of each state when the root node of the present invention has three states;

[0042] Figure 7 This invention provides a directed graph of electronic system accidents. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0044] The present invention provides a method for assessing the reliability status of wind turbine generator sets based on a directed accident graph, comprising the following steps:

[0045] S1: Establishing a directed graph model of wind turbine generator accidents based on fault tree analysis; specifically including:

[0046] S101: Study the components and failure modes of wind turbine generator sets, and build a fault tree model of wind turbine generator sets based on the components and failure modes.

[0047] The structure of a wind turbine generator set is as follows Figure 1 As shown, the wind turbine subsystem mainly includes blades, hubs, and internal pitch control devices; the gearbox system mainly includes gearboxes and main shafts; the generator electronics system includes stators, rotors, and front and rear bearings in the generator; the hydraulic subsystem includes hydraulic pumps, controllers, motors, pressure sensors, and other components; the yaw subsystem includes yaw motors, yaw bearings, yaw gears, and other components; and the nacelle and tower subsystem mainly includes the nacelle and tower.

[0048] As a specific example, a power generation system (GES) is a device that converts the mechanical energy output from a wind turbine into electrical energy through the principle of electromagnetic induction. GES operates under varying conditions and complex environments for extended periods, involving physical and chemical processes such as mechanics, electricity, magnetism, and force, making it prone to failure. A fault tree model of a power generation system is shown below. Figure 2 As shown, common faults in generator systems include stator faults, rotor faults, bearing faults, and other faults. Stator fault modes are categorized into winding faults, core faults, and base faults. Rotor fault modes include rotor support fatigue fracture due to repeated unbalanced loads and severe rotor vibration, and permanent magnet demagnetization due to severe rotor vibration, instantaneous high temperatures, and central overcurrent. Bearing faults include abnormal bearing vibration caused by excessive load, excessive clearance, and excessive vibration, as well as wear due to insufficient lubrication. In addition, generator systems may also experience other fault modes such as ventilation hood faults and cooling fan faults.

[0049] S102: Convert the logic gates in the fault tree model into directed arcs, convert the events in the fault tree model into nodes of the fault directed graph, and generate a conditional confidence quality table based on the meaning of the logic gates.

[0050] Fault tree analysis requires events to exist in only two states: normal and fault. Furthermore, it cannot represent the uncertainty of fault logic relationships between events, thus failing to meet the reliability state assessment requirements for wind turbine generators. Directed fault graphs, on the other hand, can not only address the polymorphism and uncertainty of fault logic relationships in wind turbine generators but also have a mapping relationship with fault trees. Therefore, fault tree analysis is used to establish a directed fault graph model.

[0051] From the concepts of directed accident graphs and fault trees, and the process of researching problems, it can be concluded that there is a mapping relationship between nodes in directed accident graphs and events in fault trees; both aim to identify system components, clarify research problems, and complete the description of the entire system. Therefore, they are structurally similar, and events in fault trees can be converted into nodes in directed accident graphs. Similarly, directed arcs in directed accident graphs and logic gates in fault trees both express the relationship between components and the system; therefore, logic gates can be converted into directed arcs.

[0052] S103: Combine the directed arcs, accident directed graph nodes, and conditional confidence quality table to establish an accident directed graph model for wind turbine generator sets.

[0053] Considering that a fault tree can only represent two states, namely a normal state and a fault state, while a directed fault graph can represent multiple states, namely a normal state, a fault state, and a normal state or a fault state (this state indicates that it is unknown whether the current state is a normal state or a fault state, i.e., an uncertain state), taking the fault tree of an "AND gate" and an "OR gate" as an example, assuming that there are only two states, {Yes} and {No}, the conditional reliability quality table in the directed fault graph is shown in Table 1.

[0054] Table 1. Conditional Reliability Quality Table for AND and OR Gates

[0055]

[0056] In Table 1, the AND gates define S as {Yes} only when both R1 and R2 are {Yes}. When either R1 or R2 is {No}, S is {No}. When only one of R1 and R2 is {Yes} and the other is uncertain ({Yes, No}), S is also uncertain. Similarly, in Table 1, the OR gates define S as {Yes} only when either R1 or R2 is {Yes}. When both R1 and R2 are {No}, S is {No}. When either R1 or R2 is {No} and the other is uncertain, S is also uncertain ({Yes, No}). In the conditional reliability quality table for wind turbine generators, probability values ​​are used to express the uncertain logical relationships between components and the system.

[0057] In summary, the process of establishing a directed graph model of an accident using fault tree analysis is as follows: Figure 3 As shown.

[0058] S2: Based on the state importance method, perform fault polymorphism reliability assessment on the directed graph model of the wind turbine generator set accident;

[0059] System reliability state assessment based on accident directed graphs refers to calculating the state information of leaf nodes using accident directed graph reasoning (i.e., uncertainty reasoning based on evidence theory) given the accident directed graph model and root node state information.

[0060] The model of the directed graph of the accident is as follows Figure 4 As shown, A is the root node because it has no parent node; D, E, F, and G are leaf nodes because they have no child nodes; B and C are child nodes because they have a parent node A; D and E are child nodes of B, and B is also a child node of A; F and G are child nodes of C, and C is also a child node of A.

[0061] Taking a two-level directed graph model of an accident as an example, R1 and R2 are two root nodes, and S is a leaf node. Each node has three states: normal, semi-fault, and fault, represented by 0, 1, and 2, respectively. The semi-fault state represents an uncertain state, which may be normal or faulty. Table 2 shows the confidence intervals of the root nodes in each state, where Bel(R) = ... w ), w = 0, 1, 2 is the lower bound of the interval, Pl(R) w ), w = 0, 1, 2 is the upper bound of the interval.

[0062] Table 2: Status Information of the Root Node

[0063]

[0064] According to the law of total probability, the failure probability of a leaf node in each state is:

[0065]

[0066] In the formula, w S (w S w1(w1=0,1,2) represents the state of the leaf node. w2(w2=0,1,2) represents the state of the root node R1. This indicates that leaf node S is in w S The probability of failure in a given state. This represents the probability that the root node R1 is in state w1. This represents the probability of the root node R2 being in state w2. This indicates that when root node R1 is in state w1 and root node R2 is in state w2, leaf node S is in state w. S Failure probability of a state. The root node can only exist in one state, so each state of the root node is subject to the constraints of formulas (2) and (3).

[0067]

[0068] The goal of system reliability state assessment based on accident directed graph is to obtain an interval value, that is, the confidence interval of the system in each state. Therefore, Bel (lower probability) and Pl (upper probability) of the components are used for calculation, and the obtained results are used as the interval boundary to obtain the reliability of the wind turbine generator.

[0069]

[0070] In the formula, N represents the number of root nodes (i.e., components); L represents the number of intermediate nodes (i.e., subsystems); π(Y) l ) represents the intermediate node Y l The set of parent nodes of leaf node S; π(S) represents the set of parent nodes of leaf node S.

[0071] As described in the problem description, the probabilities of each state of the root node are confidence intervals, and theoretically, any probability value can occur within the confidence interval. However, due to the constraint of the sum of the probabilities of each state of the root node, the actual acceptable range of each state is narrowed. Therefore, the key to the reliability state assessment of wind turbine generators based on the accident directed graph is to obtain the true acceptable range of each state of the root node and select appropriate points within the acceptable range for accident directed graph reasoning.

[0072] Specifically, it includes:

[0073] S201: Analyze the state of the root node in the directed graph model of the accident, and determine the importance of each state of the fault based on the structure of the directed graph of the accident and the research problem;

[0074] like Figure 5 This diagram illustrates the possible ranges of each state when the root node has two states: normal and fault. Since the initial possible ranges for each state are confidence intervals, the joint possible range is the shaded quadrilateral in the diagram, defined as the initial possible range. However, due to the constraint of the sum of the probabilities of each state, the actual possible range lies within the thick blue line in the diagram, defined as the actual possible range. That is, because of the constraint of the sum of the probabilities of each state, the possible range shrinks from within the initial quadrilateral to within the line segment. Furthermore, when the probability of a certain state is constant, the possible ranges of the other states will also shrink accordingly. When the probability of the normal state is Pl(R... 0 When ), the probability of the fault state can only take the value P2(R). 2 )=1-Pl(R 0 That is, the range of the probability of the fault state changes from the original thick blue line to a fixed value.

[0075] Similarly, when the root node has three states—normal, semi-faulty, and faulty—as... Figure 6 As shown, the initial possible range is within a spatial cuboid. Due to the constraint of the sum of probabilities of each state, the actual range of values ​​is narrowed to P(R). 0)+P(R 1 )+P(R 2 =1 and within the cross section of the initial acceptable range.

[0076] The above analysis shows that the constraint of summing the probabilities of root node states narrows the practically acceptable range. Furthermore, when the value of a certain state is fixed, the range of values ​​for the other states also changes, making it difficult to select appropriate points for directed graph propagation of the accident. Therefore, the state importance method is used to assess the reliability of wind turbine generators. The state importance of a root node refers to the degree of influence of each root node state on the system, which is related to the structure of the directed graph of the accident and the research problem. For example, when the research problem is to obtain the confidence interval of leaf nodes in a normal state, the importance of the normal state of the root node is greater than that of the fault state; in other words, the normal state has a greater impact on the system than the fault state.

[0077] S202: Sort the root node states according to their importance, and based on the constraints of the sum of the states, take the maximum value within the acceptable range to evaluate the reliability of each state of the root node.

[0078] S203: Taking all importance rankings into account, perform directed graph reasoning for wind turbine generator set accidents according to the same importance ranking, and take the maximum value from it to realize intelligent reliability state assessment of multi-state faults of wind turbine generator set and obtain the confidence interval of each state.

[0079] Since the importance of the root node's state is related to the structure of the directed graph and the research problem, and the structures of different systems often differ, it is difficult to determine the magnitude of state importance. Therefore, all rankings are considered. For example, when the root node has two states, there are two possible rankings of state importance: normal state > fault state, and fault state > normal state. When the root node has w states, there are w! possible rankings of state importance. Then, within a given ranking, the maximum value is selected from the highest to the lowest possible value.

[0080] Optionally, taking the two states of the root node as an example, the initial confidence intervals for the normal state and the fault state of the root node are respectively [Bel(R 0 ),Pl(R 0 )]、[Bel(R 2 ),Pl(R 2 When the importance of states is ranked as normal state > fault state, the normal state takes the maximum value within the initial possible range, i.e., P1(R). 0 )=Pl(R 0 Due to the limitation of the sum of the states of the root node, the range of possible fault states will be narrowed. If the fault state has a range of values, then the maximum value within that range, i.e., P1(R), will be taken. 2If no value is available for the fault state, i.e., Bel(R) 2 )>1-Pl(R 0 Then, decrease P1(R) sequentially. 0 ) takes a value until P1(R) 2 There is a possible value for P1(R), in fact, when P1(R) 2 When a value is available, the value is P1(R). 2 ) = Bel(R 2 The reason for this is that when taking the maximum value of the initial confidence interval under normal conditions, the constraint of the sum of probabilities was not considered, resulting in an excessively large value.

[0081] Then, the probabilities of normal and fault states obtained under this state importance ranking, i.e., P1 = [P1(R 0 ),P1(R 1 [], perform directed graph reasoning on the accident to obtain a set of probabilities of leaf nodes being in each state. If the normal state has any value within the initial confidence interval, and the fault state has no possible value, then it indicates that the initial confidence interval has Bel(R). 0 )+Bel(R 2 )>1 or Pl(R) 0 )+Pl(R 2 If the probability is less than 1, then there is no suitable point for directed graph reasoning in case of the accident. Similarly, when the importance of states is ranked as fault state > normal state, the probabilities of the root node's normal state and fault state can be obtained as P2 = [P2(R)]. 0 ),P2(R 2 We perform directed graph reasoning on the incident to obtain another set of probability values ​​for each state of the leaf nodes. We then take the maximum or minimum value from the two sets of results to obtain the confidence interval for each state of the leaf nodes.

[0082] Optionally, if the root node has three states, then there are 3! possible orderings for state importance. One order is chosen, for example: normal state > semi-fault state > fault state. Then, the states are ranked from most important to least important within the acceptable range, and the most important state (normal state) is selected, taking its maximum value P1(R) within the initial confidence interval. 0 )=Pl(R 0 When the probability values ​​of the half-fault state and the fault state are obtained, the root node becomes two states, except that the constraint condition changes from the original sum of probabilities being 1 to P1(R). 1 )+P1(R 2 )=1-P1(R 0 If the root node's normal state takes the maximum value P1(R) within the initial confidence interval. 0 )=Pl(R 0If no values ​​are available for the half-fault and fault states, then the values ​​for the normal states are further reduced until values ​​are available for the half-fault and fault states. This process is repeated, ranking the importance of the six states and obtaining the probabilities of each state of the root node using the above method. Then, directed graph reasoning is performed to obtain the probabilities of six leaf nodes in each state, and the maximum and minimum values ​​are taken to obtain the confidence intervals for each state of the leaf nodes.

[0083] In the embodiments, the root node can be divided into two or three states, indicating that this method is also applicable to situations where the number of root node states is inconsistent.

[0084] As a specific example, Figure 7 This is a directed graph model of an electronic system accident, with 14 root nodes R1 to R2. 14 There are 7 intermediate nodes Y1 to Y7 and one leaf node S. Each node has three states: normal state, half-fault state and fault state, represented by 0, 1 and 2 respectively.

[0085] Root node R1 indicates excessive load as the cause of rotor support fatigue fracture; root node R2 indicates severe vibration as the cause of rotor support fatigue fracture; root node R3 indicates rotor vibration as the cause of permanent magnet demagnetization; root node R4 indicates instantaneous high temperature as the cause of permanent magnet demagnetization; root node R5 indicates excessive current as the cause of permanent magnet demagnetization; root node R6 indicates excessive load as the cause of abnormal vibration in the front and rear bearings; root node R7 indicates excessive clearance as the cause of abnormal vibration in the front and rear bearings; root node R8 indicates excessive vibration as the cause of abnormal vibration in the front and rear bearings; root node R9 indicates winding fault as the cause of stator fault. 10 This indicates that the cause of the stator fault is a core fault, with the root node R. 11 Indicates the cause of the stator fault is a base fault; root node R 12 Indicates the cause of other faults: windshield failure, root node R. 13 Indicates the cause of other faults, such as fan failure; root node R 14 The intermediate node Y1 indicates insufficient lubrication as the cause of the front and rear bearing failures. Intermediate node Y2 indicates rotor support fatigue fracture, permanent magnet failure, abnormal vibration of the front and rear bearings, stator failure, rotor failure, rotor and front / rear bearing failures, and other failures. Leaf node S indicates a generator / electronics system failure.

[0086] As can be seen from the directed graph model of the electronic system accident, child nodes are connected to parent nodes through "OR gates". For example, child nodes R1 and R2 are connected to parent node Y1 through "OR gate", child nodes Y1 and Y2 are connected to parent node Y5 through "OR gate", child nodes Y4, Y5, Y6 and Y7 are connected to parent node S through "OR gate", and so on.

[0087] In this configuration, child nodes R1 and R2 are connected to parent node Y1 via an OR gate. When either R1 or R2 is in a fault state, Y1 is also in a fault state. When both R1 and R2 are in a fault-free state, Y1 is also in a fault-free state. Furthermore, when either R1 or R2 is in a fault-free state and the other is in an uncertain state, Y1 is also in an uncertain state. Other connection relationships follow the same logic.

[0088] Table 3 shows the confidence intervals for each state of the root node in the electronic system. Table 4 shows the conditional confidence quality table for intermediate node Y1. Table 5 shows the conditional confidence quality table for intermediate node Y6. Table 6 shows the conditional confidence quality table for leaf node S, representing the fault logic relationship between parent and child nodes.

[0089] Table 3. Confidence Intervals for Each State of the Root Node of the Electronic System

[0090]

[0091] Table 4 Conditional Reliability Quality Table for Intermediate Node Y1 (Rotor Support Fatigue Fracture) of the Generating Electronics System

[0092]

[0093] Table 5 Conditional Reliability Quality Table for Intermediate Node Y6 (Front and Rear Bearing Failure) of the Electronic System

[0094]

[0095]

[0096] Table 6 Conditional Reliability Quality Table of Intermediate Node S in the Electronic System

[0097]

[0098] The state importance method was used to assess the reliability of the generator electronic system of the wind turbine generator, and the confidence intervals for each state are shown in Table 7. Analysis revealed that the confidence intervals obtained by the state importance method for the generator electronic system in the normal and semi-fault states are [0.6493, 0.8680] and [0.0679, 0.1394], respectively. These intervals include [0.6927, 0.8228] and [0.0887, 0.1301] obtained by the Monte Carlo method, and are closer to the Monte Carlo method's values ​​of [0.6142, 0.8754] and [0.0531, 0.1727] than those obtained by the confidence interval method. The confidence interval for the generator system in a fault state obtained using the state importance method is [0.0642, 0.2113], which includes [0.0692, 0.1940] obtained using the Monte Carlo method. However, the confidence interval for the generator system in a fault state obtained using the confidence interval method is [0.0716, 0.2131], which does not include the confidence interval obtained using the Monte Carlo method. Therefore, the confidence interval for the generator system in each state obtained using the state importance method is more accurate than the confidence interval method.

[0099] Table 7 Confidence intervals for the electronic system in each state

[0100]

[0101] During the analysis, the Monte Carlo method was used for 10... 6 Six incidents of directed graph reasoning were performed using the state importance method, and two incidents of directed graph reasoning were performed using the confidence interval method.

[0102] Therefore, by comparing the results obtained by the state importance method of this invention with the Monte Carlo method and the confidence interval method, the effectiveness of this invention in analyzing the reliability of wind turbine generator sets is verified.

[0103] In the embodiments of the wind turbine generator reliability status assessment method based on accident directed graphs provided by the present invention, the functional units in each step can be integrated into a controller processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units. The integrated unit implemented as a software functional unit can be stored in a computer-readable storage medium. The software functional unit stored in the storage medium includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the method described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0104] The present invention also discloses a wind turbine generator reliability status assessment system based on a fault directed graph, comprising: at least one processor; and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor can execute the wind turbine generator reliability status assessment method based on a fault directed graph as described above by calling the program instructions.

[0105] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for assessing the reliability status of wind turbine generator sets based on a directed accident graph, characterized in that, include: S1: Establish a directed graph model of wind turbine generator accidents based on fault tree analysis; S101: Study the components and failure modes of wind turbine generator sets, and build a fault tree model of wind turbine generator sets based on the components and failure modes. S102: Convert the logic gates in the fault tree model into directed arcs, convert the events in the fault tree model into nodes of the fault directed graph, and generate a conditional confidence quality table based on the meaning of the logic gates. S103: Combine the directed arcs, accident directed graph nodes, and conditional confidence quality table to establish an accident directed graph model for wind turbine generator sets; S2: Based on the state importance method, perform fault polymorphism reliability assessment on the directed graph model of the wind turbine generator set.

2. The method according to claim 1, characterized in that, S2 include: S201: Analyze the state of the root node in the directed graph model of the accident, and determine the importance of each state of the fault based on the structure of the directed graph of the accident and the research problem; S202: Sort the root node states according to their importance, and based on the constraints of the sum of the states, take the maximum value within the acceptable range to evaluate the reliability of each state of the root node. S203: Taking all importance rankings into account, perform directed graph reasoning for wind turbine generator set accidents according to the same importance ranking, and take the maximum value from it to realize intelligent reliability state assessment of multi-state faults of wind turbine generator sets and obtain the confidence interval of each state.

3. The method according to claim 2, characterized in that, In S201, the number of node states includes three types: normal, half-fault, and fault.

4. The method according to claim 3, characterized in that, In the directed graph model of the wind turbine generator set accident, the child nodes are connected to the parent node through an "OR gate". That is, when any child node is in a fault state, the parent node is in a fault state; when all child nodes are in a fault-free state, the parent node is in a fault-free state; and when any child node is in an uncertain state and the other child nodes are in a fault-free state, the parent node is also in an uncertain state.

5. The method according to claim 4, characterized in that, The failure probabilities of leaf nodes in each state in the aforementioned directed graph model of the accident are: In the formula, w S (w S w1(w1=0,1,2) represents the state of the leaf node, w2(w2=0,1,2) represents the state of the root node R1, and w2(w2=0,1,2) represents the state of the root node R2. This indicates that leaf node S is in w S The probability of failure in a state. This represents the probability of the root node R1 being in state w1. This represents the probability of the root node R2 being in state w2. This indicates that when root node R1 is in state w1 and root node R2 is in state w2, leaf node S is in state w. S The probability of failure in a given state.

6. The method according to claim 5, characterized in that, The constraints for each state of the root node in the directed graph model of the accident are as follows:

7. The method according to claim 2, characterized in that, The formula for calculating the confidence interval of the system in each state in the directed graph model of the accident is as follows: In the formula, N represents the number of root nodes (i.e., components); L represents the number of intermediate nodes (i.e., subsystems); π(Y) l ) represents the intermediate node Y l The set of parent nodes of leaf node S; π(S) represents the set of parent nodes of leaf node S.

8. A reliability status assessment system for wind turbine generator sets based on a directed accident graph, characterized in that, include: At least one processor; as well as At least one memory communicatively connected to the processor, wherein: The memory stores program instructions that can be executed by the processor, which can call the program instructions to execute the wind turbine generator reliability status assessment method based on the fault directed graph as described in any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium stores computer instructions that cause the computer to execute the wind turbine generator reliability status assessment method based on a fault-directed graph as described in any one of claims 1 to 7.

Images