CN110645153B - Wind generating set fault diagnosis method and device and electronic equipment - Google Patents

Wind generating set fault diagnosis method and device and electronic equipment Download PDF

Info

Publication number
CN110645153B
CN110645153B CN201810677108.2A CN201810677108A CN110645153B CN 110645153 B CN110645153 B CN 110645153B CN 201810677108 A CN201810677108 A CN 201810677108A CN 110645153 B CN110645153 B CN 110645153B
Authority
CN
China
Prior art keywords
fault
mode
reason
failure
probability
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810677108.2A
Other languages
Chinese (zh)
Other versions
CN110645153A (en
Inventor
司伟涛
邓刚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Goldwind Science and Creation Windpower Equipment Co Ltd
Original Assignee
Beijing Goldwind Science and Creation Windpower Equipment Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Goldwind Science and Creation Windpower Equipment Co Ltd filed Critical Beijing Goldwind Science and Creation Windpower Equipment Co Ltd
Priority to CN201810677108.2A priority Critical patent/CN110645153B/en
Publication of CN110645153A publication Critical patent/CN110645153A/en
Application granted granted Critical
Publication of CN110645153B publication Critical patent/CN110645153B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics

Abstract

The embodiment of the application provides a wind generating set fault diagnosis method, a wind generating set fault diagnosis device and electronic equipment, wherein the fault diagnosis method comprises the following steps: determining a current fault mode of the wind generating set detected in real time; analyzing an FMECA fault information knowledge base according to the influence hazard of the fault mode of the wind generating set, and determining all possible fault reasons of the current fault mode, the prior probability of each fault reason and the conditional probability of the current fault mode corresponding to all fault reasons; calculating posterior probabilities of the fault reasons corresponding to the current fault mode based on the prior probabilities of the fault reasons and the conditional probability of the current fault mode; and generating a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason. Through the scheme of the embodiment of the application, the fault diagnosis capability and the diagnosis effect of the wind generating set can be effectively improved.

Description

Wind generating set fault diagnosis method and device and electronic equipment
Technical Field
The invention relates to the technical field of wind power generation, in particular to a wind generating set fault diagnosis method and device and electronic equipment.
Background
Wind power generation has been widely used around the world as a means of generating green renewable energy. As the number of the wind generating sets increases and the running time is accumulated, the probability of the wind generating sets failing is relatively increased. The performance of the wind generating set is reduced once the wind generating set fails, and even the wind power generation cannot be carried out. Therefore, the fault diagnosis of the wind generating set is an essential link for wind power generation.
All parts of the wind generating set are mutually associated and closely coupled, so that great challenges are brought to the maintenance, repair and other guarantee service work of the wind generating set. At present, in the prior art, most of the predictive diagnosis on the faults of the wind generating set adopts statistical fault history data, an information base of a fault tree structure is established according to the fault history data, and when the faults occur, related fault tree information is obtained by searching on the basis of the information base to obtain a fault reason set. However, the establishment of the fault tree needs to be based on the investigation and analysis of the accident, the workload is large, sufficient data support is needed, the post analysis is performed, the fault diagnosis effect is not ideal, the possible fault reasons need to be eliminated one by field maintenance personnel, and the fault positioning efficiency is low.
Disclosure of Invention
The embodiment of the application provides a wind generating set fault diagnosis method and device and electronic equipment, and the fault diagnosis effect can be improved. The technical scheme adopted by the embodiment of the application is as follows:
in a first aspect, an embodiment of the present application provides a wind turbine generator system fault diagnosis method, including:
determining a current fault mode of the wind generating set detected in real time;
determining all possible fault reasons of the current fault mode, prior probability of each fault reason and conditional probability of the current fault mode corresponding to all fault reasons according to a fault mode influence hazard analysis (FMECA) fault information knowledge base of a preconfigured wind generating set;
calculating posterior probabilities of the fault reasons corresponding to the current fault mode based on the prior probabilities of the fault reasons and the joint probability distribution of the current fault mode;
and generating a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason.
In a second aspect, an embodiment of the present application provides a wind generating set fault diagnosis device, including:
the fault mode determining module is used for determining the current fault mode of the wind generating set detected in real time;
the fault cause and probability determination module is used for analyzing an FMECA fault information knowledge base according to the influence harmfulness of the fault mode of the pre-configured wind generating set, and determining all possible fault causes of the current fault mode, the prior probability of each fault cause and the conditional probability of the current fault mode corresponding to all the fault causes;
the failure cause posterior probability determining module is used for calculating the posterior probability of each failure cause corresponding to the current failure mode based on the prior probability of each failure cause and the conditional probability of the current failure mode;
and the fault processing scheme generating module is used for generating a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason.
In a third aspect, an embodiment of the present application provides an electronic device, including a processor and a memory;
a memory configured to store machine-readable instructions;
and the processor is used for executing the wind generating set fault diagnosis method shown in any embodiment of the application by calling the readable instructions.
The technical scheme provided by the embodiment of the application has the following beneficial effects: according to the wind generating set fault diagnosis method, the wind generating set fault diagnosis device and the electronic equipment, fault reasons when the wind generating set is in fault are diagnosed by fully utilizing FMECA knowledge established in a product design stage of the wind generating set, and a fault processing scheme list comprising the fault reasons and posterior probabilities of the fault reasons can be obtained. Because FMECA is fault analysis data generated in the product research and development stage, continuous iteration optimization can be performed along with the deepening of cognition, the fault diagnosis can be performed by using the knowledge, the fault cause can be more effectively positioned, and particularly, the fault diagnosis capability and the fault diagnosis effect of the wind generating set are improved for the positioning of the fault causes of some new fault modes.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.
FIG. 1 is a schematic flow chart illustrating a method for diagnosing a fault of a wind turbine generator system according to an embodiment of the present application;
FIG. 2 is a schematic flow chart illustrating a method for diagnosing a fault of a wind turbine generator system according to another embodiment of the present application;
FIG. 3 is a schematic flow chart illustrating a wind turbine generator system fault diagnosis method provided in another embodiment of the present application;
FIG. 4 is a schematic flow chart illustrating a method for diagnosing a fault of a wind turbine generator system according to an exemplary embodiment of the present disclosure;
FIG. 5 is a diagram illustrating a Bayesian network in accordance with an exemplary embodiment of the present application;
FIG. 6 is a schematic diagram of another Bayesian network in accordance with an exemplary embodiment of the present application;
FIG. 7 is a schematic diagram of a Bayesian network including a posterior probability of each cause of failure when left yaw feedback is lost in an exemplary embodiment of the present application;
FIG. 8 is a diagram illustrating a Bayesian network including a posterior probability of each failure cause when the left yaw is lost and the right yaw is fed back normally in an exemplary embodiment of the present application;
fig. 9 is a schematic structural diagram illustrating a fault diagnosis apparatus for a wind turbine generator system according to an embodiment of the present application;
fig. 10 is a schematic structural diagram illustrating a fault diagnosis device for a wind generating set provided in another embodiment of the present application;
fig. 11 is a schematic structural diagram illustrating a fault diagnosis device for a wind generating set provided in another embodiment of the present application;
fig. 12 is a schematic structural diagram illustrating a wind turbine generator system fault diagnosis apparatus provided in yet another embodiment of the present application;
fig. 13 is a schematic structural diagram of an electronic device in an embodiment of the present application;
description of reference numerals in the drawings:
900-fault diagnosis device; 910-FMECA knowledge base creation module;
920-a failure mode determination module; 930-fault cause and probability determination module;
940-a fault detection event determination module; 950 — a failure cause posterior probability determination module;
960-a fault handling scheme generation module; 970-fault cause location module.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present invention.
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
The terms referred to in this application will first be introduced and explained:
FMECA: aiming at all possible faults of a product, determining the influence of each fault mode on the work of the product according to the analysis of the fault modes, finding out single-point faults, and determining the hazard of the single-point faults according to the severity and the occurrence probability of the fault modes, wherein the single-point faults are a fault analysis model generated in the research and development stage of the product.
FMECA meter of the wind generating set: the data table or document based on the FMECA of the wind turbine generator system records, but not limited to, product or function marks, functional descriptions of products or functions in the wind turbine generator system, fault modes in the wind turbine generator system, fault causes causing the fault modes, prior probabilities of the fault causes, conditional probabilities of the fault modes corresponding to the fault causes, namely, probabilities of occurrence of the fault modes under a condition that one fault cause occurs (also referred to as fault mode influence probabilities), sources of the prior probabilities of the fault causes, and severity of the fault modes, namely, degrees of influence on the wind turbine generator system.
Probability table: the probability table describes conditional probabilities for each failure mode in the FMECA table.
Fault handling measures table: the probability table describes a failure detection method (i.e., a failure cause troubleshooting plan) for each failure cause in the FMECA table.
An FMECA fault information knowledge base of the wind generating set comprises the following steps: the database is established based on an FMECA table, wherein at least information of each fault mode, each fault reason, the prior probability of each fault reason, the conditional probability of each fault mode and the like in the wind generating set is recorded, and other data in the FMECA table, data in a fault handling measure table, data calculated or deduced based on the data in the FMECA table, data increased according to the fault analysis requirement and the like can be further included in the knowledge base.
Bayesian Network (BN): the method is a probability network combining image theory and probability theory, and is a directed acyclic graph model.
Failure diagnosis Bayesian network model (FDBN): an inference model that expresses various information and their interrelations related to system failure prediction in a network structure. In the embodiment of the present application, the FDBN is a BN-based wind turbine generator system fault mode diagnosis model built step by step according to FMECA knowledge, and may be expressed by a triple < X, a, P >, where:
X={X1,X2,X3,...,Xnthe set of all nodes in the BN network represents relevant variables in the actual fault prediction problem;
A={[aij]i is more than or equal to 1 and less than or equal to n, j is more than or equal to 1 and less than or equal to n, is a directed edge set connecting all nodes and represents the incidence relation among all variables, aijRepresenting a slave node XiTo node XjA directed edge of (a);
P={P(Xi|π(Xi)),Xie.X is the node X in the networkiThe associated probability distribution, the strength of the connection between the expression variables,represents XiIs selected. XiFor the bottom layer reason, P ═ P (X)i|π(Xi)),Xie.X is the prior probability of the node, XiWhen not the bottom-most cause, P ═ { P (X)i|π(Xi)),XiE X is the conditional probability of the node.
Most of the existing wind generating set fault diagnosis schemes are realized based on historical statistical data, a large amount of data is needed for supporting, the fault diagnosis effect is poor due to post analysis, the obtained fault reason diagnosis sets need to be eliminated one by field personnel, and the fault positioning efficiency is poor.
The application provides a wind generating set fault diagnosis method, a wind generating set fault diagnosis device and electronic equipment, and aims to solve the technical problems in the prior art.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Fig. 1 shows a flow chart of a wind generating set fault diagnosis method provided by an embodiment of the present application. As shown in fig. 1, the fault diagnosis method may mainly include the following steps:
step S102: determining a current fault mode of the wind generating set detected in real time;
the failure mode, i.e., the manifestation of a failure, is typically a specification describing the occurrence of a failure phenomenon that can be observed or measured by the product. For the wind generating set, real-time fault detection can be carried out on all equipment, components and the like of the wind generating set according to actual requirements so as to obtain a detected current fault mode. For example, for a yaw control loop of a wind park, fault modes typically include, but are not limited to, loss of left yaw feedback, loss of right yaw feedback, inability of yaw control contactor contacts to close, and the like.
It should be noted that, the detection of the fault mode may be completed by using a fault detection system of an existing wind turbine generator system, or may be a system that integrates the fault detection system with the fault diagnosis device of the embodiment of the present application and has the functions of fault detection and fault diagnosis. That is to say, the determination of the current failure mode may be implemented by obtaining a real-time detection result of a single failure detection system, or may be implemented by detecting and determining a system itself having both failure detection and failure diagnosis functions, and the determination manner of the current failure mode in the embodiment of the present application is not limited.
Step S103: determining all possible fault reasons of a current fault mode, prior probability of each fault reason and conditional probability of the current fault mode corresponding to all fault reasons according to an FMECA fault information knowledge base of the wind generating set;
the FMECA fault information knowledge base is a corresponding relation database which is created based on an FMECA table of the wind generating set and comprises all fault modes, all fault reasons, the prior probability of all fault reasons and the conditional probability of all fault modes of the wind generating set. After the current fault mode is determined, all possible fault reasons of the current fault mode can be searched in the FMECA fault information knowledge base based on the current fault mode, and the prior probability of each fault reason and the conditional probability of the current fault mode are further determined.
In the embodiment of the application, for a certain fault reason, the prior probability of the fault reason is the fault rate of the fault reason, and the fault rate of a product represents the actual fault rate of an analyzed product in a task stage of the product. The conditional probability of the failure mode refers to the conditional probability of all direct causes corresponding to the failure mode (i.e., all parent nodes of the failure mode), where when there are multiple direct causes corresponding to the failure mode, the conditional probability of the failure mode is the conditional probability under the joint distribution of the multiple direct causes, and when there is only one direct cause of the failure mode, the conditional probability of the failure mode is the conditional probability of the failure mode under the condition that the one direct cause occurs.
Step S104: calculating posterior probabilities of the fault reasons corresponding to the current fault mode based on the prior probabilities of the fault reasons and the conditional probability of the current fault mode;
as can be seen from the foregoing description, the conditional probability of the current failure mode is the conditional probability of the joint distribution of all failure causes of the failure mode. The posterior probability of each failure cause corresponding to the current failure mode refers to the probability of each failure cause occurring when the current failure mode has occurred. For a fault reason of the current fault mode, when the prior probability of each fault reason of the current fault mode and the conditional probability of the current fault mode are known, the posterior probability of each fault reason corresponding to the current fault mode can be calculated and obtained according to a Bayesian formula respectively.
The prior probability of a certain fault cause is the actual fault rate of the fault cause, for example, the loose left yaw connection is a fault cause of a left yaw loss feedback fault mode, and the fault rate of the loose left yaw connection is 1.20E-2, namely the prior probability.
Step S105: and obtaining a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason.
According to the wind generating set fault diagnosis method provided by the embodiment of the application, the FMECA knowledge established in the product design stage of the wind generating set is fully utilized to realize the diagnosis of the fault reasons when the wind generating set has faults, and the fault processing scheme list comprising the fault reasons and the posterior probability of the fault reasons can be obtained. Because FMECA is fault analysis data generated in the product research and development stage, continuous iteration optimization can be performed along with the deepening of cognition, the fault diagnosis can be performed by using the knowledge, the fault cause can be more effectively positioned, and particularly, the fault diagnosis capability and the fault diagnosis effect of the wind generating set are improved for the positioning of the fault causes of some new fault modes.
In the embodiment of the present application, all possible failure reasons of the current failure mode may be all possible direct reasons of the current failure mode, or may be all possible lowest layer reasons of the current failure mode.
The direct cause is a failure cause directly causing the current failure mode to occur. The bottommost failure cause, i.e., the bottommost failure cause, is a failure cause for which no other failure cause is present in the failure mode, and the bottommost failure cause is also a bottommost failure mode, and in addition to the failure of the bottommost failure cause, no failure cause is present for which the failure causes.
That is to say, in the embodiment of the present application, the depth of the fault diagnosis may be located to the direct cause causing the fault mode to occur, or may be located to the lowest cause causing the fault mode to occur. When the fault reason is a direct reason, the posterior probability of each direct reason and each direct reason is included in the fault processing scheme list, and when the fault reason is the lowest layer reason, the posterior probability of each lowest layer reason and each lowest layer reason is included in the fault processing scheme list.
As can be seen from the foregoing description, the lowest-level cause refers to the lowest-level cause of the failure mode, and the lowest-level cause may be the direct cause of the failure mode or may be the further lower-level cause of the failure mode. It can be seen that the failure cause may be a cause of occurrence of a certain failure mode, or may be a failure mode caused by other failure causes.
For example, loss of left yaw feedback for a failure mode may be a direct cause of its occurrence, possibly due to loose left yaw wiring, failure of left yaw control contactors, and so forth. Since there is no fault cause causing the looseness of the left yaw connection, the looseness of the left yaw connection is the cause of the fault of the bottommost layer. And for the failure of the left yaw control contactor, the failure cause may be the failure of the main contact of the left yaw control contactor to attract, or the failure of the auxiliary contact to attract, that is, the failure of the left yaw control contactor is not only a direct cause of the loss of the left yaw feedback, but also a failure mode caused by the failure of the main contact to attract or the failure of the auxiliary contact to attract.
In an embodiment of the application, as shown in fig. 2, the wind turbine generator system fault diagnosis method may further include:
step S101: and creating an FMECA fault information knowledge base of the wind generating set.
In the embodiment of the present application, creating the FMECA fault information knowledge base of the wind turbine generator system may specifically include:
counting the conditional probability of each fault mode in an FMECA table;
and creating an FMECA fault information knowledge base according to the corresponding relation among each fault mode, each fault reason, the prior probability of each fault reason and the statistical conditional probability of each fault mode in the FMECA table.
In this embodiment of the application, when creating the FMECA fault information knowledge base of the wind turbine generator system, the creating may further include:
and determining a troubleshooting scheme of each fault reason in an FMECA table of the wind generating set.
Correspondingly, the creating an FMECA fault information knowledge base according to the correspondence between the fault modes, the fault reasons, the prior probabilities of the fault reasons, and the statistical conditional probabilities of the fault modes in the FMECA table may specifically include:
and establishing an FMECA fault information knowledge base according to the corresponding relation among each fault mode, each fault reason, the prior probability of each fault reason, the statistical conditional probability of each fault mode and the determined fault troubleshooting scheme of each fault reason in the FMECA table.
The troubleshooting scheme of the fault cause is a detection method/diagnosis troubleshooting measure of the fault cause. Correspondingly, the fault processing scheme list may also include a troubleshooting scheme for each fault cause of the current fault mode, so that field personnel can quickly troubleshoot the fault cause based on the troubleshooting scheme in the list.
In practical application, when the FMECA fault information knowledge base is established, a fault troubleshooting scheme, which is a detection tool and a detection method corresponding to each fault reason in the FMECA table, can be determined according to the FMECA table of the wind generating set, a fault handling measure table is generated, a probability table is generated by counting the conditional probability of each fault mode in the FMECA table, and then the FMECA table, the fault handling measure table and the probability table are summarized to obtain the FMECA fault information knowledge base of the wind generating set.
It can be understood that the FMECA fault information knowledge base is only required to be created once in the initial stage, and in the subsequent application process, data in the knowledge base can be continuously optimized in various ways such as deepening of cognition, learning of operation data, expert correction and the like.
In this embodiment of the present application, after determining all possible failure reasons of the current failure mode, the method may further include:
determining fault detection events corresponding to fault reasons according to an FMECA fault information knowledge base;
if a fault detection event corresponding to a fault cause exists, calculating a posterior probability of each fault cause corresponding to the current fault mode based on the prior probability of each fault cause and the conditional probability of the current fault mode corresponding to all fault causes may specifically include:
determining the state of a fault detection event corresponding to a fault reason;
and calculating the posterior probability of each fault reason according to the detection state of the fault detection event corresponding to the fault reason, the prior probability of each fault reason, the conditional probability of the current fault mode and the conditional probability of the fault detection event.
The fault detection event refers to the uppermost fault mode except the current fault mode caused by the fault reason, and the uppermost fault mode refers to the fault mode which does not cause other fault modes to occur, that is, the uppermost fault mode is not the fault reason of other fault modes. For example, for a failure mode, the PLC module 140DI9 (serial number of PLC module) fails to have a failure cause, i.e., a failure Logic Controller (PLC) module 140DI9, which is also one of the possible causes of failure of the right yaw feedback, and the right yaw feedback loss does not have a failure mode of the previous layer caused by the failure feedback loss, so that the right yaw feedback failure is the top layer failure mode, and the right yaw feedback failure is a failure detection event corresponding to the failure cause, i.e., the failure PLC module 140DI9 failure.
In calculating the posterior probability of each failure cause, the failure detection event corresponding to the failure cause can be further considered, and since the state of the failure detection event can be determined by detection, the state of the failure detection event is determined, and the posterior probability of each failure cause is calculated based on the state of the failure detection event, so that the accuracy of the posterior probability calculation of the failure cause can be improved. The state of the fault detection event refers to the event occurring or not occurring.
Fig. 3 shows a schematic flow chart of a fault diagnosis method for a wind generating set provided in another embodiment of the present application. As can be seen from fig. 3, the fault diagnosis method shown in fig. 3 may further include, on the basis of the fault diagnosis method shown in fig. 1, after obtaining the fault handling scheme list of the current fault mode:
step S106: and sequentially carrying out fault troubleshooting on all fault reasons according to the fault processing scheme list until the actual fault reason of the current fault mode is positioned.
If the fault processing scheme list comprises fault troubleshooting schemes of all fault reasons, field maintenance personnel can directly troubleshoot according to the troubleshooting schemes in the list. If no corresponding troubleshooting scheme exists in the list, field personnel can perform troubleshooting on the failure reason according to experience.
In the embodiment of the present application, troubleshooting is performed on each failure reason in sequence according to the failure processing scheme list, which may specifically include:
troubleshooting a fault reason in the fault processing scheme list, and determining the state of the fault reason;
if the state of one fault reason is the fault, determining whether the fault reason is the actual fault reason of the current fault mode;
and if the state of one fault reason is that the fault does not occur, determining that the posterior probability of one fault reason is zero, updating the fault processing scheme list, and continuously performing fault troubleshooting on other fault reasons according to the updated fault processing scheme list.
Specifically, when the state of one fault cause is that a fault does not occur, the posterior probability of the fault cause in the fault handling scheme list may be directly updated to 0, the posterior probabilities of other fault causes in the list may be recalculated based on the updated posterior probability, the fault handling scheme list may be updated according to the recalculated result, and then another fault cause may be checked based on the updated fault handling scheme list. By the method, the on-site maintenance personnel can quickly and accurately locate the actual fault reason, and the actual fault locating efficiency is improved.
In the embodiment of the present application, generating the fault handling scheme list of the current fault mode according to each fault cause and the posterior probability of each fault cause may specifically include:
and according to the sequence of the posterior probability of each fault reason from large to small, creating a fault processing scheme list comprising each fault reason and the posterior probability of each fault reason.
Correspondingly, troubleshooting is performed on each fault reason in sequence according to the fault handling scheme list, which may specifically include:
and sequentially troubleshooting the fault reasons according to the sequence of the posterior probability of the fault reasons from large to small.
By generating the fault processing scheme list according to the sequence of the posterior probability of each fault reason from large to small, field personnel can conveniently perform troubleshooting from the first fault reason in the list, namely the fault reason with the maximum probability, and quickly find the reason which most possibly causes the fault.
In this embodiment of the application, if all possible failure causes of the current failure mode are direct causes, the determining, according to the FMECA failure information knowledge base, all possible failure causes of the current failure mode, the prior probability of each failure cause, and the conditional probability of the current failure mode corresponding to all failure causes may specifically include:
searching all possible direct reasons of the current fault mode in an FMECA fault information knowledge base;
establishing a Bayesian network with directed edges pointing to the fault mode from the fault reason according to the cause-effect relationship of the fault by taking the current fault mode as the fault mode node and the found direct reasons as the fault reason nodes;
determining prior probability of each fault reason node and conditional probability of a fault mode node in the Bayesian network according to an FMECA fault information knowledge base;
at this time, correspondingly, based on the prior probability of each fault cause and the conditional probability of the current fault mode, calculating the posterior probability of each fault cause corresponding to the current fault mode may specifically include:
and calculating the posterior probability of each fault reason node corresponding to the fault mode node according to the prior probability of each fault reason node and the conditional probability of the fault mode node.
It can be understood that the prior probability of the failure cause node and the conditional probability of the failure mode node are the prior probability of the corresponding failure cause and the conditional probability of the failure mode.
In the embodiment of the present application, if all possible failure causes of the current failure mode are all possible bottom-layer causes of the current failure mode, determining, according to an FMECA failure information knowledge base of the wind turbine generator system, all possible failure causes of the current failure mode, a prior probability of each failure cause, and a conditional probability of the current failure mode corresponding to all failure causes, which may specifically include:
searching all possible direct reasons of the current fault mode in an FMECA fault information knowledge base;
determining whether the direct cause is the bottom cause of the current failure mode;
if the direct reason section is not the bottom-layer reason, further searching the corresponding bottom-layer reason in the FMECA fault information knowledge base by taking the direct reason as a fault mode;
establishing a Bayesian network with directed edges pointing to a fault mode from fault reasons according to a fault cause-and-effect relationship by taking the current fault mode as a fault mode node, all searched bottom layer reasons (including direct reasons which are the bottom layer reasons and bottom layer reasons which are further searched according to direct reasons which are not the bottom layer reasons) as fault reason nodes, and searched fault reasons except the bottom layer reasons as intermediate nodes;
determining prior probability of each fault reason node, prior probability of each intermediate node, conditional probability of each intermediate node and conditional probability of a fault mode node in the Bayesian network according to an FMECA fault information knowledge base; correspondingly, calculating the posterior probability of each failure cause node corresponding to the failure mode node based on the prior probability of each failure cause node and the conditional probability of the failure mode node may specifically include:
and calculating the posterior probability of each fault reason node corresponding to the fault mode node based on the prior probability of each fault reason node, the prior probability of each intermediate node, the conditional probability of each intermediate node and the conditional probability of the fault mode node in the Bayesian network.
According to the embodiment of the application, an FDBN model containing the corresponding relation between the current fault mode and each fault reason (direct reason or bottommost reason) of the current fault mode can be obtained by establishing a Bayesian network, and based on the model, field maintenance personnel or fault analysis can visually see the causal relation of various related information of the current fault mode, and the posterior probability of each fault reason can be directly displayed in the model.
In this embodiment of the application, if there is a corresponding fault detection event for the lowest-layer reason, the bayesian network may further include a fault detection node, and a directed edge pointed to the corresponding fault reason node by the fault detection node, where the fault detection node is the fault detection event corresponding to the lowest-layer reason.
As can be seen from the foregoing description, the failure detection event is the uppermost failure mode except the current failure mode caused by the failure cause, and therefore, the failure detection node is the node corresponding to the uppermost failure mode except the current failure mode caused by the failure cause.
By adding the fault detection nodes in the Bayesian network, the Bayesian network is more perfected, a more finished FDBN model is obtained, and the addition of the fault detection nodes can further improve the accuracy of fault cause positioning and improve the efficiency of fault cause positioning as can be known from the foregoing description.
As can be seen from the foregoing description, when there is a fault detection node and the posterior probability of each fault cause is calculated, the conditional probability of the fault detection event needs to be determined according to the FMECA fault information knowledge base.
For better illustration and understanding of the wind turbine generator system fault diagnosis method provided by the present application, the method is further described in detail below with reference to an embodiment.
In this embodiment, the current fault mode is described by taking the left yaw feedback loss in the yaw control loop of the wind turbine generator system as an example, and the depth of location of the fault cause in this embodiment is the bottommost cause. Fig. 4 shows a schematic flow chart of the wind turbine generator system fault diagnosis method in this embodiment, which specifically includes:
firstly, an FMECA fault information knowledge base is established, as shown in fig. 4, the way of establishing the knowledge base in this embodiment is as follows: determining detection tools and detection methods (troubleshooting schemes) corresponding to fault reasons in an FMECA (failure mode-matching) table, and generating a fault processing measure table; counting the conditional probability of each fault mode in an FMECA (failure mode event correlation) table to obtain a probability table; and summarizing the fault processing measure table, the FMECA table and the probability table to obtain an FMECA fault information knowledge base.
Part of the data for the yaw control loop portion summarized by the FMECA table and the fault handling measures table in this particular embodiment is shown in table 1. The left yaw control relay 110K10, the left yaw control contactor 110K6, the PLC module 140DI9, the left yaw control relay 110K12 and the left yaw control contactor 110K8 shown in table 1 are all devices in the yaw control circuit of the wind turbine generator system in this embodiment, wherein 110K10, 110K6, 140DI9, 110K12 and 110K8 are numbers of the devices.
TABLE 1
Table 2 shows part of data of the yaw control loop section in the probability table in this embodiment. The failure cause column shows all the direct causes of the failure modes, and the conditional probability column shows the conditional probability of each failure mode. For example, p (M | C)1C2C3C4) 1 indicates that under the condition that the four failure causes, left yaw control relay 110K10 failure, left yaw tie loosening, left yaw control contactor 110K6 failure, and PLC module 140DI9 failure occur simultaneously, the conditional probability that the left yaw feedback loses the failure mode,then a conditional probability of left yaw feedback loss is indicated under the condition that left yaw control relay 110K10 failure, left yaw tie loose, and left yaw control contactor 110K6 failure occurred, and PLC module 140DI9 failure did not occur.
TABLE 2
As can be seen from tables 1 and 2, in the FMECA fault information knowledge base in this embodiment, the data includes, but is not limited to, product or function flags of the wind turbine generator system, a function description (shown in a function column in the table), a fault mode, a fault cause of the fault mode, a severity of the fault cause (the severity is higher, the fault cause is described to have a larger influence on the wind turbine generator system), a fault detection method (i.e., a troubleshooting plan of the fault cause), a prior probability of the fault cause (shown in a fault rate column), and a conditional probability of each fault mode.
It should be noted that, in the present embodiment, each table in the FMECA failure information knowledge base is presented in a form of a list, in an actual application, the specific form of the FMECA failure information knowledge base may be formulated according to actual needs, and may include but is not limited to the form shown in the present embodiment, and the content of the data included in the FMECA failure information knowledge base may also be adjusted according to needs. The FMECA table fault handling measure table and the probability table may be combined into one table or may be three separate tables.
Any row of records in the FMECA table (which may be a table obtained by summarizing the FMECA table and the fault handling measure table, or an FMECA table without summarizing the fault handling measure table) may be referred to as an FMECA information unit. The FMECA information unit represents various related information about a certain fault mode, fault reasons, fault influence, a fault detection method and the like, and is the information basis of fault diagnosis. For example, the record in row 1 in table 1 is an information unit in which a plurality of pieces of related information of the failure mode of the left yaw loss feedback are described, and the failure mode probability in the information unit is 0.8, which indicates that when a single failure cause that the left yaw control relay 110K10 fails occurs, the probability that the left yaw feedback is lost is 80%.
And after the FMECA fault information knowledge base is established, fault diagnosis can be carried out on the basis of the knowledge base. The fault detection system based on the wind generating set detects the wind generating set in real time, loads fault characteristic information (current fault mode) when a fault is triggered, and realizes fault diagnosis based on the fault characteristic information detected in real time and an FMECA fault information knowledge base.
In this specific embodiment, the fault characteristic information is left yaw feedback loss, when it is determined that a fault mode of left yaw feedback loss occurs, the left yaw feedback loss is used as a fault mode event, that is, the fault mode node M searches all fault causes using the left yaw feedback loss as a fault mode, that is, direct causes of the fault mode, in the FMECA fault information knowledge base, further searches next-layer fault causes of the fault mode using the direct fault causes in the FMECA fault information knowledge base according to all the searched direct causes, and sequentially searches downward until the searched fault causes are the lowest-layer causes. The searched lowest-layer cause is taken as a fault cause event (fault cause node C in the present embodiment), and the remaining fault causes are taken as intermediate events (intermediate nodes D in the present embodiment).
Based on the data shown in Table 1, all direct causes searched for failure modes with loss of left yaw feedback include four failure causes, left yaw control relay 110K10 failure, left yaw line wiring slack, left yaw control contactor 110K6 failure, and PLC module 140DI9 failure. Based on these four failure causes, the data shown in table 1 is further searched for the next layer failure cause with these four failure causes as the failure mode, and then two failure causes, namely, failure cause of main contact failure and failure cause of auxiliary contact failure in row 21 and row 22 in table 1 can be searched for, and these two failure causes are the next layer failure cause with failure mode of left yaw control contactor 110K 6. By adopting the mode, downward searching is performed according to the hierarchy until the searched failure reason is the reason of the lowest layer.
Based on the data shown in fig. 1, the bottom layer is due to failure of the left yaw control relay 110K10, loose left yaw wiring, failure of the PLC module 140DI9, failure of the main contacts to engage and failure of the auxiliary contacts to engage. Taking these five bottom-most causes as the fault cause node C, the left yaw control contactor 110K6 failed as the middle node D, and directed edges were sent from the fault cause to the fault pattern to indicate causal relationships, creating a bayesian network, as shown in fig. 5. In fig. 5, M represents a fault mode node, D1 corresponds to an intermediate node, and C1 to C5 correspond to the five fault cause nodes, respectively, where C1 indicates that the left yaw control relay 110K10 fails, C2 indicates that the left yaw connection is loose, C5 indicates that the PLC module 140DI9 fails, C3 indicates that the main contact fails, and C4 indicates that the auxiliary contact fails.
After the searching of the bottommost layer reasons is completed, on the basis of all fault reason nodes, further searching fault detection events corresponding to the bottommost layer reasons in an FMECA fault information knowledge base, namely further searching all fault modes which are not searched yet and take the fault reasons as fault reasons to the upper layer, sequentially searching to the upper layer until the searched fault modes have no fault reasons of other information units matched with the fault reasons, namely the searched fault modes are the topmost layer fault modes, adding the searched fault detection events as fault detection nodes E into the Bayesian network shown in the figure 5, and then sending directed edges from the fault reason nodes to the fault detection nodes to indicate a relationship in a cause-and-effect manner to obtain a further improved Bayesian network.
In this embodiment, based on the detected failure cause node, that is, failure of the left yaw control relay 110K10, loosening of the left yaw connection, failure of the PLC module 140DI9, failure of the main contact, and failure of the auxiliary contact, a search is further performed in the FMECA failure information repository shown in table 1 to the upper layer, it can be found that the failure mode right yaw feedback loss is the failure cause of the PLC module 140DI9, and then the right yaw feedback loss is used as the failure detection node E and added to the bayesian network in fig. 5 to obtain the bayesian network shown in fig. 6, and the bayesian network is used as the FDBN in this embodiment.
It is understood that the node identifiers C1-C5 in fig. 5 and 6 do not correspond exactly to the fault cause identifiers C1-C4 in table 2 above. The respective failure cause identifications of each row in table 2 indicate the respective failure causes of the failure mode of the row.
After the FDBN is established, the prior probability of the corresponding fault reason node and the intermediate node, the conditional probability of the intermediate node and the conditional probability of the fault mode node in the established FDBN are determined based on an FMECA fault information knowledge base, and on the basis, the posterior probability of each fault reason node is calculated through Bayesian formula and according to the state inference of the fault mode node.
Specifically, if the failure detection node E is not considered, the posterior probability of each node may be calculated for the nodes C1, C2, and C5 by using the bayesian formula according to the prior probability of each node and the conditional probability of the node M (the conditional probability of the 1 st row in table 2). For the nodes C3 and C4, the posterior probability of the node D may be calculated by the bayesian formula according to the prior probability of the node D and the conditional probability of the node M, and the posterior probabilities of the nodes C3 and C4 may be calculated according to the posterior probability of the node D, the prior probabilities of the nodes C3 and C4, and the conditional probability of the node D (the conditional probability of the 2 nd row in table 2).
Table 3 shows the posterior probability of each failure cause node when the node E is not considered in this embodiment. A schematic diagram of a bayesian network corresponding to the posterior probabilities in table 3 is shown in fig. 7, where Y corresponding to node M shown as 100% (100%) (corresponding N is 0) indicates that the current failure mode occurs, Y indicates the posterior probability when the corresponding failure cause occurs, and N indicates the probability of non-occurrence, such as the posterior probability P (C1) of nodes C1-C5, respectively1I M) is 0.82%, i.e., when it represents that the left yaw feedback is lost, the probability of failure of the left yaw control relay 110K10 is 0.0082.
TABLE 3
Posterior probability Cause of failure Diagnostic and troubleshooting measures
p(C2|M)=0.226 Left yawing connection looseness Hand-pulled visual observation
p(C3|M)=0.0383 The main contact of the left yaw control contactor 110K6 can not be attracted Taking a multimeter to measure
p(C4|M)=0.0165 The auxiliary contact of the left yaw control contactor 110K6 can not be pulled in Taking a multimeter to measure
p(C1|M)=0.0082 Left yaw control relay 110K10 failure Taking a multimeter to measure
p(C5|M)=0.0054 PLC Module 140DI9 failure Taking a multimeter to measure
The posterior probability P (C) is illustrated by the left yaw line slack or node C2 in Table 32| M) is explained. The Bayesian formula shows that:
wherein:
the sign between two probabilities in the above formula represents the multiplication of the two probabilities. With p (C)1) For example, p (C)1) Indicating a prior probability of the occurrence of a failure of left yaw control relay 110K10,
if the fault detection node E is considered, the posterior probability of the nodes C1 to C5 under the condition that the left yaw feedback is lost and the right yaw feedback state is determined is calculated through a Bayes formula according to the prior probability of the nodes C1 to C5, the prior probability of the node D, the conditional probability of the node M and the conditional probability of the node E.
In this embodiment, the state of the node E is determined by performing right yaw operation, and after performing the right yaw operation, it is assumed that no right yaw feedback is lost, that is, the state of the node E is not generated, and is recorded as state 0. The posterior probabilities of the nodes C1 through C5 are calculated based on the state according to the node E and the state of the node M.
The posterior probabilities of C1 through C5 calculated from the state of node E and the state of node M are shown in Table 4. Fig. 8 shows a schematic diagram of a bayesian network corresponding to the posterior probability in table 4. As can be seen from table 4 and fig. 8, the posterior probability calculated at this time is the posterior probability of the failure cause in the normal mode of left yaw feedback loss and right yaw feedback, such as the posterior probability p (C) of the node C22I ME) ═ 0.227, it means that when left yaw feedback is lost and right yaw feedback is normal, the posterior probability of loosening left yaw connections is 0.227, and since the state of right yaw feedback is also determined, the posterior probability in table 4 is more accurate than that in table 3.
TABLE 4
After the calculation of the posterior probability of each fault reason node is completed, all triggered bottom events (the lowest fault reason events) are sorted from the great to the small of the posterior probability, and the diagnosis and troubleshooting method is displayed to form a fault processing scheme list. Table 3 or table 4 shows a list of failure handling schemes in this specific embodiment.
Based on table 4, the field maintenance personnel can check the 1 st fault reason according to the diagnosis and checking measures shown in the table, and feed back the checking result to the diagnosis system, so that the system recalculates the posterior probability of each bottommost reason according to the checking and feedback result, updates the fault processing scheme list, and sequentially updates until the fault reason is positioned to solve the fault. Specifically, according to the generated fault handling scheme list, firstly, the looseness of the left yaw connection is checked and eliminated, if the connection is not loosened, the posterior probability corresponding to the looseness of the left yaw connection can be updated to 0 according to the checking result, on the basis, the posterior probabilities of other fault reasons are updated to obtain an updated fault handling scheme list, on-site maintenance personnel check the fault reason with the highest probability in the new list based on the new fault handling scheme list, and the actual fault reason is accurately and efficiently located through the operation of fault checking and updating the fault handling scheme after the fault checking.
Corresponding to the wind generating set fault diagnosis method shown in fig. 1, the embodiment of the application also provides a wind generating set fault diagnosis device. As shown in fig. 9, the fault diagnosis apparatus 900 may include a fault mode determination module 920, a fault cause and probability determination module 930, a fault cause posterior probability determination module 950, and a fault handling scheme generation module 960. Wherein:
a failure mode determination module 920, configured to determine a current failure mode of the wind turbine generator system detected in real time;
a failure cause and probability determination module 930, configured to determine, according to a FMECA failure information knowledge base of a preconfigured wind turbine generator system, all possible failure causes of a current failure mode, a prior probability of each failure cause, and a conditional probability of the current failure mode corresponding to all failure causes;
a failure cause posterior probability determination module 950, configured to calculate posterior probabilities of the failure causes corresponding to the current failure mode based on the prior probabilities of the failure causes and the conditional probability of the current failure mode;
a failure handling scheme generating module 960, configured to generate a failure handling scheme list of the current failure mode according to each failure cause and the posterior probability of each failure cause.
In the embodiment of the application, all possible fault reasons of the current fault mode are direct reasons of the current fault mode or the lowest layer reasons of the current fault mode; the direct cause is a failure cause directly causing the occurrence of the current failure mode, and the lowest failure cause is a failure cause which is a failure cause for which no failure cause has occurred in the failure mode.
The wind generating set fault diagnosis device provided by the embodiment of the application realizes the diagnosis of the fault reason when the wind generating set fault occurs by utilizing the FMECA knowledge established in the product design stage of the wind generating set, and can obtain the fault processing scheme list comprising each fault reason and the posterior probability of each fault reason. Because FMECA is fault analysis data generated in the product research and development stage, fault diagnosis can be performed in advance by using the knowledge, the fault cause can be more effectively positioned, and especially the fault causes of some new fault modes are positioned, so that the fault diagnosis capability and the fault diagnosis effect of the wind generating set are improved.
It is understood that the modules of the fault diagnosing apparatus 900 shown in fig. 9 may correspond to the steps of the fault diagnosing method shown in fig. 1, and have a function of executing the corresponding method steps. For detailed functional description of the fault diagnosis apparatus shown in fig. 9, reference may be made to the above detailed description of the fault diagnosis method shown in fig. 1, and details are not repeated here.
In the embodiment of the present application, the fault diagnosis apparatus 900 may further include a fault detection event determination module 940, as shown in fig. 10.
The failure detection event determining module 940 is configured to determine a failure detection event corresponding to each failure cause according to the FMECA failure information knowledge base after determining all possible failure causes of the current failure mode, where the failure detection event refers to a top-level failure mode except the current failure mode caused by the failure cause.
Correspondingly, if a fault detection event corresponding to the fault cause exists, the posterior probability of fault cause determination module 950 may be specifically configured to:
determining the detection state of a fault detection event corresponding to a fault reason;
and calculating the posterior probability of each fault reason according to the detection state of the fault detection event corresponding to the fault reason, the prior probability of each fault reason, the conditional probability of the current fault mode and the conditional probability of the fault detection event.
In this embodiment, the fault diagnosis apparatus 900 may further include a fault cause location module 970, as shown in fig. 11.
And the fault cause positioning module 970 is configured to perform fault troubleshooting on each fault cause in sequence according to the fault processing scheme list until an actual fault cause of the current fault mode is located.
In this embodiment of the application, when the fault cause location module 970 performs fault troubleshooting on each fault cause in sequence according to the fault handling scheme list, it may specifically be configured to:
troubleshooting a fault reason in the fault processing scheme list, and determining the state of the fault reason;
if the state of one fault reason is the fault, determining whether the fault reason is the actual fault reason of the current fault mode;
and if the state of one fault reason is that the fault does not occur, determining that the posterior probability of one fault reason is zero, updating the fault processing scheme list, and continuously performing fault troubleshooting on other fault reasons according to the updated fault processing scheme list.
In this embodiment of the present application, the fault handling scheme generating module 960 may be specifically configured to:
according to the sequence of the posterior probability of each fault reason from big to small, a fault processing scheme list comprising each fault reason and the posterior probability of each fault reason is created;
when the fault cause location module 970 performs fault troubleshooting on each fault cause in sequence according to the fault handling scheme list, it may specifically be configured to:
and sequentially troubleshooting the fault reasons according to the sequence of the posterior probability of the fault reasons from large to small.
In this embodiment of the application, if the failure cause is the lowest cause of the current failure mode, the failure cause and probability determining module 930 may be specifically configured to:
searching all possible direct reasons of the current fault mode in an FMECA fault information knowledge base;
determining whether the direct cause is the bottom cause of the current failure mode;
if the direct reason section is not the bottom-layer reason, further searching the corresponding bottom-layer reason in the FMECA fault information knowledge base by taking the direct reason as a fault mode;
establishing a Bayesian network with directed edges pointing to the fault mode from the fault reason according to the cause-effect relationship of the fault by taking the current fault mode as the fault mode node, all the searched bottom layer reasons as the fault reason nodes and the searched fault reasons except the bottom layer reasons as intermediate nodes;
and determining the prior probability of each fault reason node, the prior probability of each intermediate node, the conditional probability of each intermediate node and the conditional probability of the fault mode node in the Bayesian network according to the FMECA fault information knowledge base.
Accordingly, the posterior probability of failure cause determination module 950 can be specifically configured to: and calculating the posterior probability of each fault reason node corresponding to the fault mode node based on the prior probability of each fault reason node, the prior probability of each intermediate node, the conditional probability of each intermediate node and the conditional probability of the fault mode node in the Bayesian network.
In the embodiment of the application, if the corresponding fault detection event exists in the lowest-layer reason, the bayesian network further includes a node using the fault detection event as a fault detection node, and a directed edge pointed to the corresponding fault reason node by the fault detection node.
In the embodiment of the present application, the fault diagnosis apparatus 900 may further include an FMECA knowledge base creation module 910, as shown in fig. 12.
And the FMECA knowledge base creating module 910 is used for creating an FMECA fault information knowledge base of the wind generating set.
In this embodiment of the present application, the FMECA knowledge base creating module 910 may be specifically configured to:
counting the conditional probability of each fault mode in an FMECA table;
and creating an FMECA fault information knowledge base according to the corresponding relation among each fault mode, each fault reason, the prior probability of each fault reason and the counted conditional probability of each fault mode in the FMECA table.
In this embodiment, the FMECA knowledge base creating module 910 may further be configured to determine a troubleshooting scheme for each failure cause in the FMECA table of the wind turbine generator system.
Correspondingly, the FMECA knowledge base creating module 910 may specifically be configured to, when creating the FMECA fault information knowledge base according to the correspondence between the fault modes, the fault reasons, the prior probabilities of the fault reasons, and the statistical conditional probabilities of the fault modes in the FMECA table:
and establishing an FMECA fault information knowledge base according to the corresponding relation among each fault mode, each fault reason, the prior probability of each fault reason, the statistical conditional probability of each fault mode and the fault troubleshooting scheme of each fault reason in the FMECA table.
In this embodiment, the fault handling scheme list may further include a troubleshooting scheme for each fault reason.
It is understood that each module of the wind generating set fault diagnosis device in the embodiment of fig. 9 and based on fig. 9 may have a function of implementing the corresponding step of the wind generating set fault diagnosis method in the embodiment of fig. 1 or based on fig. 1. The function can be realized by hardware, and can also be realized by executing corresponding software by hardware. The modules can be software and/or hardware, and can be implemented individually or by integrating a plurality of modules. For the functional description of each module of the wind generating set fault diagnosis device, reference may be made to fig. 1 or the corresponding description in the wind generating set fault diagnosis method based on fig. 1, and details are not repeated here.
An embodiment of the present application further provides an electronic device, as shown in fig. 13, an electronic device 2000 shown in fig. 13 includes: a processor 2001 and a transceiver 2004. The processor 2001 is coupled to the transceiver 2004, such as via the bus 2002. Optionally, the electronic device 2000 may further include a memory 2003. It should be noted that the transceiver 2004 is not limited to one in practical applications, and the structure of the electronic device 2000 is not limited to the embodiment of the present application.
Among them, the processor 2001 is applied to the embodiment of the present application to realize the functions of the failure diagnosis apparatus 900 shown in fig. 9 to 12. The transceiver 2004 includes a receiver and a transmitter, and the transceiver 2004 may be applied to the embodiment of the present application to realize the function of the fault diagnosis apparatus 900 of the embodiment of the present application to communicate with other systems/apparatuses/modules in the wind turbine.
The processor 2001 may be a CPU, general purpose processor, DSP, ASIC, FPGA or other programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 2001 may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs and microprocessors, and the like.
Bus 2002 may include a path that conveys information between the aforementioned components. The bus 2002 may be a PCI bus or an EISA bus, etc. The bus 2002 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in FIG. 13, but this is not intended to represent only one bus or type of bus.
The memory 2003 may be, but is not limited to, a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, an EEPROM, a CD-ROM or other optical disk storage, optical disk storage (including compact disk, laser disk, optical disk, digital versatile disk, blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
Optionally, the memory 2003 is used for storing application program code for performing the disclosed aspects, and is controlled in execution by the processor 2001. The processor 2001 is configured to execute application program codes stored in the memory 2003 to implement the operations of the respective modules in the failure diagnosis apparatus 900 according to the embodiment of the present application.
The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims (14)

1. A wind generating set fault diagnosis method is characterized by comprising the following steps:
determining a current fault mode of the wind generating set detected in real time;
determining all possible fault reasons of the current fault mode, the prior probability of each fault reason and the conditional probability of the current fault mode corresponding to all the fault reasons according to a preset fault mode influence hazard analysis (FMECA) fault information knowledge base of the wind generating set;
calculating posterior probabilities of the fault reasons corresponding to the current fault mode based on the prior probabilities of the fault reasons and the conditional probability of the current fault mode;
generating a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason; the fault reason is a direct reason of the current fault mode or a bottommost reason of the current fault mode;
the direct cause is a fault cause directly causing the current fault mode to occur, and the bottommost fault cause is a fault cause which is a fault cause that does not exist when the fault mode is the fault mode.
2. The method of claim 1, wherein after determining all possible causes of the fault in the current fault mode, the method further comprises:
determining a fault detection event corresponding to each fault reason according to the FMECA fault information knowledge base, wherein the fault detection event refers to the uppermost fault mode except the current fault mode caused by the fault reason;
if a fault detection event corresponding to the fault reason exists, calculating a posterior probability of each fault reason corresponding to the current fault mode based on the prior probability of each fault reason and the conditional probability of the current fault mode corresponding to all the fault reasons, including:
determining the state of a fault detection event corresponding to the fault reason;
and calculating the posterior probability of each fault reason according to the detection state of the fault detection event corresponding to the fault reason, the prior probability of each fault reason, the conditional probability of the current fault mode and the conditional probability of the fault detection event.
3. The method of claim 2, wherein if the failure cause is the lowest cause of the current failure mode, the determining all possible failure causes of the current failure mode, the prior probability of each failure cause, and the conditional probability of the current failure mode corresponding to all the failure causes according to the FMECA failure information knowledge base of the wind turbine generator system comprises:
searching all possible direct reasons of the current fault mode in the FMECA fault information knowledge base;
determining whether the direct cause is a bottom-most cause of the current failure mode;
if the direct reason section is not the lowest layer reason, further searching the corresponding lowest layer reason in the FMECA fault information knowledge base by taking the direct reason as a fault mode;
establishing a Bayesian network with directed edges pointing to the fault mode from the fault reason according to the cause-effect relationship of the fault by taking the current fault mode as the fault mode node, all the searched bottom layer reasons as fault reason nodes and the searched fault reasons except the bottom layer reasons as intermediate nodes;
determining prior probability of each fault reason node, prior probability of each intermediate node, conditional probability of each intermediate node and conditional probability of the fault mode node in the Bayesian network according to the FMECA fault information knowledge base;
calculating a posterior probability of each of the failure causes corresponding to the current failure mode based on the prior probability of each of the failure causes and the conditional probability of the current failure mode, including:
calculating a posterior probability of each of the failure cause nodes corresponding to the failure mode node based on the prior probability of each of the failure cause nodes in the Bayesian network, the prior probability of each of the intermediate nodes, the conditional probability of each of the intermediate nodes, and the conditional probability of the failure mode node.
4. The method according to claim 3, wherein if there is a corresponding fault detection event for the lowest-level cause, the bayesian network further comprises a fault detection node and a directed edge pointed to by the fault detection node by the corresponding fault cause node, wherein the fault detection node is the fault detection event corresponding to the lowest-level cause.
5. The fault diagnosis method according to any one of claims 1 to 4, wherein after generating the fault handling scenario list of the current fault mode according to each of the fault causes and the posterior probability of each of the fault causes, further comprising:
and sequentially carrying out fault troubleshooting on each fault reason according to the fault processing scheme list until the actual fault reason of the current fault mode is positioned.
6. The method according to claim 5, wherein the generating a fault handling scenario list of the current fault mode according to each of the fault causes and the posterior probability of each of the fault causes comprises:
according to the sequence of the posterior probability of each fault reason from big to small, the fault processing scheme list comprising each fault reason and the posterior probability of each fault reason is created;
the sequentially troubleshooting each fault reason according to the fault processing scheme list comprises:
and sequentially carrying out fault troubleshooting on the fault reasons according to the sequence that the posterior probability of each fault reason is reduced from high to low.
7. The fault diagnosis method according to claim 6, wherein said sequentially troubleshooting each of the fault causes according to the fault handling scheme list comprises:
troubleshooting one fault reason in the fault processing scheme list, and determining the state of the fault reason;
if the state of the fault reason is fault, determining that the fault reason is the actual fault reason of the current fault mode;
and if the state of the fault reason is that the fault does not occur, determining that the posterior probability of the fault reason is zero, updating the fault processing scheme list, and continuously performing fault troubleshooting on other fault reasons according to the updated fault processing scheme list.
8. The fault diagnosis method according to any one of claims 1 to 4, characterized in that the method further comprises: creating the FMECA fault information knowledge base, comprising:
counting the conditional probability of each fault mode in the FMECA table;
and creating the FMECA fault information knowledge base according to the corresponding relation among the fault modes, the fault reasons, the prior probability of the fault reasons and the statistical conditional probability of the fault modes in the FMECA table.
9. A wind generating set fault diagnosis device is characterized by comprising:
the fault mode determining module is used for determining the current fault mode of the wind generating set detected in real time;
the fault cause and probability determination module is used for determining all possible fault causes of the current fault mode, the prior probability of each fault cause and the conditional probability of the current fault mode corresponding to all the fault causes according to a fault mode influence hazard analysis FMECA fault information knowledge base of the wind generating set which is pre-configured; the fault reason is a direct reason of the current fault mode or a bottommost reason of the current fault mode; the direct reason refers to a fault reason directly causing the current fault mode to occur, and the bottommost fault reason refers to a fault reason which is used as a fault mode and has no fault reason causing the fault mode to occur;
a failure cause posterior probability determination module, configured to calculate posterior probabilities of the failure causes corresponding to the current failure mode based on the prior probabilities of the failure causes and the conditional probability of the current failure mode;
and the fault processing scheme generating module is used for generating a fault processing scheme list of the current fault mode according to each fault reason and the posterior probability of each fault reason.
10. The fault diagnosis device according to claim 9, characterized in that the device further comprises:
a failure detection event determining module, configured to determine, after determining all possible failure causes of the current failure mode, a failure detection event corresponding to each failure cause according to the FMECA failure information knowledge base, where the failure detection event refers to a top-level failure mode except the current failure mode caused by the failure cause;
if the fault detection event corresponding to the fault cause exists, the fault cause posterior probability determination module is specifically configured to:
determining the state of a fault detection event corresponding to the fault reason;
and calculating the posterior probability of each fault reason according to the detection state of the fault detection event corresponding to the fault reason, the prior probability of each fault reason, the conditional probability of the current fault mode and the conditional probability of the fault detection event.
11. The apparatus according to claim 10, wherein if the failure cause is a lowest cause of the current failure mode, the failure cause and probability determination module is specifically configured to:
searching all possible direct reasons of the current fault mode in the FMECA fault information knowledge base;
determining whether the direct cause is a bottom-most cause of the current failure mode;
if the direct reason section is not the lowest layer reason, further searching the corresponding lowest layer reason in the FMECA fault information knowledge base by taking the direct reason as a fault mode;
establishing a Bayesian network with directed edges pointing to the fault mode from the fault reason according to the cause-effect relationship of the fault by taking the current fault mode as the fault mode node, all the searched bottom layer reasons as fault reason nodes and the searched fault reasons except the bottom layer reasons as intermediate nodes;
determining prior probability of each fault reason node, prior probability of each intermediate node, conditional probability of each intermediate node and conditional probability of the fault mode node in the Bayesian network according to the FMECA fault information knowledge base;
the failure cause posterior probability determination module is specifically configured to:
calculating a posterior probability of each of the failure cause nodes corresponding to the failure mode node based on the prior probability of each of the failure cause nodes in the Bayesian network, the prior probability of each of the intermediate nodes, the conditional probability of each of the intermediate nodes, and the conditional probability of the failure mode node.
12. The fault diagnosis device according to any one of claims 9 to 11, characterized in that the device further comprises:
and the fault reason positioning module is used for sequentially carrying out fault troubleshooting on each fault reason according to the fault processing scheme list until the actual fault reason of the current fault mode is positioned.
13. The apparatus according to claim 12, wherein the fault cause positioning module, when sequentially performing fault troubleshooting on each fault cause according to the fault handling scheme list, is specifically configured to:
troubleshooting one fault reason in the fault processing scheme list, and determining the state of the fault reason;
if the state of the fault reason is fault occurrence, determining whether the fault reason is an actual fault reason of the current fault mode;
and if the state of the fault reason is that the fault does not occur, determining that the posterior probability of the fault reason is zero, updating the fault processing scheme list, and continuously performing fault troubleshooting on other fault reasons according to the updated fault processing scheme list.
14. An electronic device comprising a processor and a memory;
the memory configured to store machine-readable instructions;
the processor is used for executing the wind generating set fault diagnosis method according to any one of claims 1 to 8 by calling the readable instructions.
CN201810677108.2A 2018-06-27 2018-06-27 Wind generating set fault diagnosis method and device and electronic equipment Active CN110645153B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810677108.2A CN110645153B (en) 2018-06-27 2018-06-27 Wind generating set fault diagnosis method and device and electronic equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810677108.2A CN110645153B (en) 2018-06-27 2018-06-27 Wind generating set fault diagnosis method and device and electronic equipment

Publications (2)

Publication Number Publication Date
CN110645153A CN110645153A (en) 2020-01-03
CN110645153B true CN110645153B (en) 2020-11-24

Family

ID=68988741

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810677108.2A Active CN110645153B (en) 2018-06-27 2018-06-27 Wind generating set fault diagnosis method and device and electronic equipment

Country Status (1)

Country Link
CN (1) CN110645153B (en)

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20120063634A (en) * 2010-12-08 2012-06-18 (주)에스엠인스트루먼트 Condition monitoring apparatus and condition monitoring method for machinery system
WO2013162090A1 (en) * 2012-04-25 2013-10-31 전자부품연구원 Middleware for controlling failure endurance of wind power generation control system
ES2538403B1 (en) * 2013-12-18 2016-04-13 Wind Inertia Technologies, S.L. System and procedure for maintenance management of industrial equipment
CN104564542B (en) * 2015-01-30 2018-10-16 上海电机学院 A kind of fault diagnosis system and its diagnostic method based on big data technology
CN106050580B (en) * 2016-08-17 2018-10-23 国电联合动力技术有限公司 A kind of driving chain of wind generating set fault diagnosis method and system
CN106640548B (en) * 2016-12-19 2019-03-29 北京金风科创风电设备有限公司 State monitoring method and device for wind power generating set
CN107563069A (en) * 2017-09-06 2018-01-09 国电联合动力技术有限公司 A kind of wind power generating set intelligent fault diagnosis method
CN107797537A (en) * 2017-11-10 2018-03-13 上海第二工业大学 A kind of prognostic and health management method applied to automatic production line
CN108133311A (en) * 2017-12-14 2018-06-08 长春工程学院 A kind of Wind turbines fault mode risk assessment and analysis method for reliability

Also Published As

Publication number Publication date
CN110645153A (en) 2020-01-03

Similar Documents

Publication Publication Date Title
CN103870659B (en) A kind of fault of numerical control machine tool analyzes method
CN104137078A (en) Operation management device, operation management method, and program
CN102998996A (en) Airborne real-time fault diagnosis method
CN104252401B (en) Weight based device status judgment method and system thereof
CN1976160A (en) A large electric system vulnerable line identifying method
CN102207998B (en) Method for evaluating service life and reliability of product based on zero-failure data
CN103500378A (en) Method and system for managing equipment defects by fault codes
Kharchenko et al. Availability assessment of computer systems described by stiff Markov chains: Case study
Tao et al. An approach to regression test selection based on hierarchical slicing technique
CN110645153B (en) Wind generating set fault diagnosis method and device and electronic equipment
Shen et al. System failure analysis based on DEMATEL-ISM and FMECA
CN109597752A (en) Fault propagation path simulation method based on complex network model
CN105528296A (en) Project-software-oriented class cluster test method
CN106874525B (en) Method and device for screening and counting equipment faults of wind turbine generator
CN110389892A (en) A kind of fault filling method based on cloud platform historical failure data
Nadi et al. Draca: Decision support for root cause analysis and change impact analysis for cmdbs
CN103365290A (en) Signal abnormality analysis based hidden fault diagnosis method for generator control system
Nijhawan et al. On development of change point based generalized SRGM for software with multiple releases
TW201913412A (zh) 故障診斷系統及方法
WO2012057656A2 (en) Method and system for constructing a model of disrupted operation of a technical object and a machine-readable medium
CN107102543A (en) The forming method and device of a kind of energy router anti-interference controller
Pózna et al. Colored petri net based diagnosis of process systems
CN105719062A (en) Method for assessing risks and weak links of power grid, with double fault probability characteristics considered
Gaikwad et al. Finding Proneness of S/W using Class Hierarchy Method
CN109032872B (en) Bayesian network-based equipment fault diagnosis method and system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant