CN116384487A - Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station - Google Patents
Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station Download PDFInfo
- Publication number
- CN116384487A CN116384487A CN202310286990.9A CN202310286990A CN116384487A CN 116384487 A CN116384487 A CN 116384487A CN 202310286990 A CN202310286990 A CN 202310286990A CN 116384487 A CN116384487 A CN 116384487A
- Authority
- CN
- China
- Prior art keywords
- entity
- fault
- class
- data
- knowledge
- 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.)
- Pending
Links
- 238000004146 energy storage Methods 0.000 title claims abstract description 111
- 238000004458 analytical method Methods 0.000 title claims abstract description 59
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 title claims abstract description 42
- 229910001416 lithium ion Inorganic materials 0.000 title claims abstract description 42
- 238000003745 diagnosis Methods 0.000 title claims abstract description 37
- 238000010276 construction Methods 0.000 title claims abstract description 36
- 238000000605 extraction Methods 0.000 claims abstract description 35
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 claims abstract description 25
- 229910052744 lithium Inorganic materials 0.000 claims abstract description 25
- 230000009471 action Effects 0.000 claims description 32
- 238000007726 management method Methods 0.000 claims description 27
- 239000013598 vector Substances 0.000 claims description 26
- 238000000034 method Methods 0.000 claims description 25
- 230000007246 mechanism Effects 0.000 claims description 21
- 238000012549 training Methods 0.000 claims description 19
- 238000012423 maintenance Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 14
- 238000004140 cleaning Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000012795 verification Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 239000003792 electrolyte Substances 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 4
- 238000012916 structural analysis Methods 0.000 claims description 4
- 238000013024 troubleshooting Methods 0.000 claims description 4
- 238000002360 preparation method Methods 0.000 claims description 3
- 238000013075 data extraction Methods 0.000 abstract description 3
- 230000000875 corresponding effect Effects 0.000 description 21
- 238000010248 power generation Methods 0.000 description 7
- 230000032683 aging Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 2
- 230000033228 biological regulation Effects 0.000 description 2
- 229910052799 carbon Inorganic materials 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000011230 binding agent Substances 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 210000001787 dendrite Anatomy 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000011068 loading method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012015 optical character recognition Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000007774 positive electrode material Substances 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000007086 side reaction Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/355—Class or cluster creation or modification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Biology (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Computational Linguistics (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Animal Behavior & Ethology (AREA)
- Software Systems (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The invention discloses a knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of an energy storage station, which fills the blank of fault diagnosis and treatment plan basic data of a large-scale energy storage lithium battery; by extracting the fault characteristics of the fault data, the ontology class and the inter-class relation of the knowledge graph are defined, and the structured data extraction mode has good expandability and self-adaptation capability, so that the knowledge extraction requirement of multi-source heterogeneous data can be met; the multi-source heterogeneous information data of the energy storage station can be fused, and finally a knowledge graph for diagnosing and analyzing the faults of the lithium ion battery of the energy storage station is formed; based on the knowledge graph, potential relation between the working parameters and faults can be found, and a knowledge system is formed, so that the rapid diagnosis and accurate treatment of the faults of the lithium battery of the energy storage station are realized.
Description
Technical Field
The invention relates to the technical field of knowledge graph construction, in particular to a knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of energy storage stations.
Background
In order to improve climate change problems and reduce carbon emissions, new energy based on renewable energy is developed on a large scale in the context of dual carbon targets, and installed quantities of photovoltaic and wind power are increasing year by year. Because photovoltaic power generation and wind power generation are greatly influenced by weather variation, the output process has randomness and volatility, wind power even has anti-peak regulation characteristics, load requirements cannot be met often, and in addition, the traditional thermal unit power generation is required to meet supply and demand balance, so that the problems of difficult grid connection, high wind discarding and light discarding cost and the like are caused, and the running cost of coal-fired power generation is increased.
By configuring energy storage on the power generation side, the problems of unpredictable photovoltaic power generation light rejection and wind power can be solved by utilizing energy time shifting and load tracking smooth output of the energy storage, and dynamic unbalance of power generation and power consumption caused by difficult prediction of a load end can be avoided, so that frequent start and stop of a traditional unit are reduced, and the overall operation economy of a power system is improved.
Due to the development of energy storage technology, the lithium ion battery is widely used for energy storage, and energy storage faults are increasingly complex. The existing fault diagnosis of the energy storage station mainly judges the fault type based on a threshold value by monitoring an external limited parameter, but causes the fault of the battery body and the fault of the battery management system to be various and complex, often not caused by one of the causes, but the phenomenon of mutual coupling among multiple causes exists, the symptom of the fault is not single, and the internal and external parameters of the battery body need to be detected to carry out comprehensive judgment; in addition, when unexpected faults occur, the fault reasons and the treatment schemes cannot be determined in time, namely, a specific expert is required to conduct analysis and treatment, so that the maintenance is not timely, the maintenance cost is higher and higher, the maintenance progress can be delayed, potential safety hazards are caused, and even life and property of staff are threatened.
The energy storage station can now collect data from 3 aspects: firstly, abundant internal and external operation data of a battery can be collected by utilizing the existing sensing and monitoring technology and are structured data; secondly, fault handling logs are semi-structured data; unstructured data in the field of fault analysis experts of energy storage batteries.
How to fuse the heterogeneous information of the multiple sources, find out the potential relation between the working parameters and faults and form a knowledge system, so that the rapid diagnosis and accurate response of the faults of the lithium battery of the energy storage station are realized, and the problem to be solved is urgent.
Disclosure of Invention
The invention mainly aims to provide a knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of an energy storage station, which aims to solve the problems of how to fuse multi-source heterogeneous information data of the energy storage station, find potential relation between working parameters and faults and form a knowledge system, thereby realizing rapid diagnosis and accurate response of the faults of the lithium ion batteries of the energy storage station.
The technical scheme provided by the invention is as follows:
a knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of energy storage stations comprises the following steps:
acquiring fault data generated in the operation process of the energy storage station;
extracting fault characteristics of the fault data, thereby defining the relationship between the ontology class and the class of the knowledge graph;
extracting the entity and the relation of the fault data based on the relation between the ontology class and the class to obtain triple data;
constructing a knowledge graph based on the triplet data;
and storing the constructed knowledge graph in a graph database management system to realize the storage and management of the knowledge graph.
Preferably, the obtaining fault data generated during the operation of the energy storage station includes:
acquiring structured service data of historical operation in an energy storage station and a semi-structured fault maintenance log, and cleaning the service data and the maintenance log;
and acquiring unstructured energy storage battery fault analysis expert domain knowledge and rules, and cleaning data to obtain specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action schemes.
Preferably, the extracting the fault characteristics of the fault data, so as to define the ontology class and the relationship between classes of the knowledge graph, includes:
extracting hardware knowledge, fault type knowledge and fault mechanism knowledge of lithium battery equipment according to the acquired fault data, abstracting and defining an ontology class of a knowledge graph, and determining a relation among classes, wherein the ontology class comprises: energy storage cabin class, single battery class, parallel module class, battery pack class, fault class, rule class, action class, value class, voltage class, internal resistance class, internal temperature class, working temperature class, cycle number class, state of charge class, gas class, use duration class, electrolyte internal resistance class and self-discharge rate class; the inter-class relationships include: the method comprises the steps of inclusion, attribute, feature, belongings, subjects or objects, presence or preparation, threshold judgment, alarm level, judgment result, mechanism or reason analysis.
Preferably, the extracting the entity and the relation of the fault data based on the relation between the ontology class and the class to obtain triple data includes:
extracting historical operation business data in the energy storage station to obtain position and operation state information;
and extracting and calculating historical operation service data in the energy storage station to acquire battery operation state data, so as to acquire the information entity of the energy storage equipment and each index type entity, and creating the information entity and each index type entity as a structured knowledge triplet.
Preferably, the extracting the entity and the relation of the fault data based on the relation between the ontology class and the class to obtain triple data further includes:
the method comprises the steps of obtaining keywords by adopting manual definition or an automatic learning mode from a corpus of reference texts, and extracting knowledge of a fault maintenance record log in a keyword extraction mode through an automatic mode or a rule extraction method.
Preferably, the expert domain knowledge and rules for fault analysis of the energy storage battery comprise thesis book related content and expert knowledge text; the entity and the relation of the fault data are extracted based on the relation between the ontology class and the class so as to obtain triple data, and the method further comprises the following steps:
acquiring the related content of paper books and expert knowledge texts, and cleaning data to obtain specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action schemes of the lithium battery, thereby obtaining a target corpus;
labeling part of data in the target corpus in an element-label mode;
distinguishing the target corpus by using specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action scheme 5 entities;
randomly dividing the marked partial data into a training set, a testing set and a verification set according to a preset proportion, and taking the rest unmarked data in the target corpus as a prediction set;
mapping the marked training set into a numerical value through a pre-training language model to obtain a word vector;
establishing a loss function according to the sequence tag information to construct an entity tag extraction model;
inputting the word vector into the label extraction model so as to extract the label to obtain the sequence entity label information, and estimating the generalization error through the verification set in the training process to update the super-parameters;
evaluating the constructed entity tag extraction model by using the test set, reconstructing the entity tag extraction model if the evaluation result is lower than a preset target, and terminating the construction of the entity tag extraction model if the evaluation result is greater than or equal to the preset target;
and inputting the rest unlabeled data in the target corpus into an entity tag extraction model to obtain sequence entity tag information, and obtaining the corresponding entity type by using a tag decoder.
Preferably, the constructing a knowledge graph based on the triplet data includes:
constructing a first knowledge graph based on the triple data corresponding to the structured historical operation business data in the energy storage station;
constructing a second knowledge graph based on the triplet data corresponding to the semi-structured troubleshooting log;
constructing a third knowledge graph based on the unstructured energy storage battery fault analysis expert domain knowledge and the triplet data corresponding to the rule;
vectorizing the first knowledge-graph, the second knowledge-graph and the third knowledge-graph through a TransD model.
Preferably, the first knowledge-graph, the second knowledge-graph and the third knowledge-graph are all expressed in the form of triples (h, r, t), wherein h represents a head entity, t represents a tail entity and r represents an entity relationship; vectorizing the first knowledge-graph, the second knowledge-graph and the third knowledge-graph through a TransD model comprises the following steps:
based on a given correct triplet D + = { (h, r, t) }, initialize entity and relationship embedding;
from D + The set Z of positive facts is obtained to generate 2 empty sets
And->
For each positive sample τ in Z + ＝(h + ,r + ,t + ) Generating a negative sample τ according to equation (1) - ＝(h - ,r - ,t - )；
Update set B + ＝B + ∪{τ + Sum B - ＝B - ∪{τ - }；
Training a TransD model by using the obtained positive sample and negative sample, adjusting parameters by adopting a gradient descent strategy, and embedding an updated entity and relationship into { (h, r, t) };
updating the gradient of the loss function and judging whether the gradient is approximately 0;
if the gradient is approximately 0, the set Z is selected again, otherwise the output entity and the relation are embedded, wherein the formula (1) is as follows:
the loss function is:
wherein τ= (h, r, t) is D + ∪D - Is a training sample of (a); if (h, r, t) ∈D + Then the following is satisfied: y is hrt =1; if (h, r, t) ∈D - Then the following is satisfied: y is hrt -1; thereby ensuring that the score of the positive sample is higher than that of the negative sample; f (f) r (h, t) is an objective function of the TransD model, and satisfies:
wherein w is r 、w h And w t The mapping vectors are additionally introduced by the TransD model, h is a head entity, t is a tail entity, r is an entity relationship, and I is a unit vector.
Preferably, if the gradient is approximately 0, the set Z is reselected, otherwise, the output entity and the relationship are embedded, and then the method further comprises:
acquiring representation vectors of the first knowledge graph, the second knowledge graph and the third knowledge graph through a trained TransD model;
based on the obtained entity vector, the similarity of the head entity and the tail entity and the similarity of neighbor nodes of the head entity and the tail entity are considered, and the similarity obtained through calculation of the formula (2) is matched with the head entity and the tail entity, so that knowledge map fusion is realized, wherein the formula (2) is as follows:
sim(A,B)＝α×sim Atrr (A,B)+(1-α)×sim NB (A,B)，
wherein A is the representing vector of the head entity, and B is the representing vector of the tail entity; sim (sim) Atrr (A, B) is the similarity of the head entity and the tail entity itself; sim (sim) NB (a, B) is the similarity of neighbor nodes of the head entity and the tail entity;
based on the head entity h and the entity relation r, calculating the tail entity t by using a trained TransD model, thereby completing the completion of the knowledge graph.
Preferably, the storing the constructed knowledge graph in the Neo4j graph database management system to realize the storage and management of the knowledge graph includes:
and storing the constructed knowledge graph in a Neo4j graph database management system, and updating the database at regular time by using increase and decrease management, informatization searching, structural analysis and decision reasoning.
Through the technical scheme, the following beneficial effects can be realized:
the knowledge graph construction method for the fault diagnosis and analysis of the lithium ion battery of the energy storage station fills the blank of the basic data of the large-scale energy storage lithium battery fault diagnosis and treatment plan; by extracting the fault characteristics of the fault data, the ontology class and the inter-class relation of the knowledge graph are defined, and the structured data extraction mode has good expandability and self-adaptation capability, so that the knowledge extraction requirement of multi-source heterogeneous data can be met; the multi-source heterogeneous information data of the energy storage station can be fused, and finally a knowledge graph for diagnosing and analyzing the faults of the lithium ion battery of the energy storage station is formed; based on the knowledge graph, potential relation between the working parameters and faults can be found, and a knowledge system is formed, so that the rapid diagnosis and accurate treatment of the faults of the lithium battery of the energy storage station are realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to the structures shown in these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a first embodiment of a knowledge graph construction method for fault diagnosis and analysis of a lithium ion battery in an energy storage station according to the present invention.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The invention provides a knowledge graph construction method for fault diagnosis and analysis of a lithium ion battery of an energy storage station.
As shown in fig. 1, in a first embodiment of a knowledge graph construction method for fault diagnosis and analysis of a lithium ion battery in an energy storage station according to the present invention, the embodiment includes the following steps:
step S110: and acquiring fault data generated in the operation process of the energy storage station.
Specifically, the fault data includes structured service data of historical operation inside the energy storage station, semi-structured fault maintenance log and unstructured expert domain knowledge and rules of fault analysis of the energy storage battery.
Step S120: and extracting fault characteristics of the fault data so as to define ontology classes and inter-class relations of the knowledge graph.
Step S130: and extracting the entity and the relation of the fault data based on the relation between the ontology class and the class to obtain the triplet data.
Step S140: and constructing a knowledge graph based on the triplet data.
Specifically, knowledge-graph fusion is constructed based on the triplet data. The knowledge graph is the knowledge graph of the fault diagnosis and analysis of the lithium ion battery of the energy storage station.
Step S150: and storing the constructed knowledge graph in a Neo4j graph database management system to realize the storage and management of the knowledge graph.
The knowledge graph construction method for the fault diagnosis and analysis of the lithium ion battery of the energy storage station fills the blank of the basic data of the large-scale energy storage lithium battery fault diagnosis and treatment plan; the entity identification work with higher accuracy can be realized under the condition of a small quantity of labels, and the problems that the entity identification is difficult to be carried out on small sample data in the field of energy storage lithium batteries and the accuracy is low are solved; by extracting the fault characteristics of the fault data, the ontology class and the inter-class relation of the knowledge graph are defined, and the structured data extraction mode has good expandability and self-adaptation capability, so that the knowledge extraction requirement of multi-source heterogeneous data can be met; the multi-source heterogeneous information data of the energy storage station can be fused, and finally a knowledge graph for diagnosing and analyzing the faults of the lithium ion battery of the energy storage station is formed; the potential relation between the working parameters and faults can be found based on the knowledge graph, so that a knowledge system is formed, the rapid diagnosis and accurate treatment of the faults of the lithium battery of the energy storage station are realized, and a favorable support is provided for the accurate and efficient retrieval of future fault diagnosis and treatment plans of the lithium battery.
In a second embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in energy storage stations according to the present invention, based on the first embodiment, step S110 includes the following steps:
step S210: and acquiring the structured service data of the internal historical operation of the energy storage station and a semi-structured fault maintenance log, and cleaning the service data and the maintenance log.
Specifically, the structured historical operation business data in the energy storage station is derived from the related sensing information and calculation data collected by the actual energy storage station and stored in the database, and the data of the related equipment devices on the energy storage station account are generally presented by a table and a standard database format, so that the structuring degree is high.
The semi-structured troubleshooting log is recorded in official text, where picture-like format files, as well as other format files, are converted in this application into text form using optical character recognition techniques.
Step S220: and acquiring unstructured energy storage battery fault analysis expert domain knowledge and rules, and cleaning data to obtain specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action schemes.
Specifically, the method is based on expert-defined modes or rules in the field, and has the advantages of fine quality and high extraction accuracy. Unstructured energy storage battery fault analysis expert domain knowledge and rules mainly come from journal papers and books of lithium battery energy storage station fault research, and expert domain knowledge or rules about battery system fault feature analysis are obtained by access domain experts or knowledge engineers, data such as specific fault types, fault phenomenon related data parameters, mechanism cause analysis, disposal action modes and the like of the lithium battery are recorded after cleaning, and the data are recorded in a text mode.
In a third embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in energy storage stations according to the present invention, based on the second embodiment, step S120 includes the following steps:
step S310: extracting hardware knowledge, fault type knowledge and fault mechanism knowledge of lithium battery equipment according to the acquired fault data, abstracting and defining an ontology class of a knowledge graph, and determining a relation among classes, wherein the ontology class comprises: energy storage cabin class, single battery class, parallel module class, battery pack class, fault class, rule class, action class, value class, voltage class, internal resistance class, internal temperature class, working temperature class, cycle number class, state of charge (SOC) class, gas class, service life class, electrolyte internal resistance class and self-discharge rate class; the inter-class relationships include: the method comprises the steps of inclusion, attribute, feature, belongings, subjects or objects, presence or preparation, threshold judgment, alarm level, judgment result, mechanism or reason analysis.
Specifically, all ontology classes and specific connection modes of relationships among the classes construct a knowledge graph through Python and py2Neo and utilize Neo4j to carry out visual display.
The energy storage cabins, the single batteries, the parallel connection modules and the battery packs are all energy storage cabins of the energy storage station, battery packs with different numbers of the same energy storage cabin, parallel connection modules and single batteries; the fault type is the fault type of hardware equipment of different layers; the rule class is a fault discrimination rule based on expert domain knowledge/rule abstraction, such as determining a threshold range by a salient feature parameter so as to discriminate fault types and fault levels; the action type runs a specific regulation mode according to the fault type and grade, if the temperature of the battery is higher due to the fact that the fan is not started, the fan is started to radiate heat; the value class is the quantification of the corresponding action, such as adjusting the fan to three gears according to the battery over-temperature problem, wherein the three gears are the values; other indexes such as voltage class are input into the rule class, and the basis of fault judgment is carried out; the subject or object in the inter-class relationship is the language of the parameter index adjusted during the action, such as increasing the working voltage, where the working voltage is the object, and the subject is similar; judging the relation of faults by judging the threshold value as a corresponding rule; the alarm level is a fault level determined by the rule; judging the fault type as a judging result; the mechanism or reason analysis is the reason for the fault; other relationships between classes are corresponding to the belonging relationships.
In a fourth embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station according to the present invention, based on the third embodiment, step S130 includes the following steps:
specifically, the historical operation business data in the energy storage station reflects the structured data collected by various sensors and measurement computing devices and stored in a standard database when the lithium battery is operated.
Step S410: and extracting historical operation business data in the energy storage station to obtain position and operation state information.
Step S420: and extracting and calculating historical operation service data in the energy storage station to acquire battery operation state data, so as to acquire the information entity of the energy storage equipment and each index type entity, and creating the information entity and each index type entity as a structured knowledge triplet.
Specifically, in this embodiment, the historical operation service data in the energy storage station is as follows in table 1:
TABLE 1
The data in table 1 is directly extracted and constructed into a dictionary structure, keys are corresponding IDs, values are entity types and inter-class relations in the table, after traversing the whole table, the entity types and the inter-class relations in the dictionary are connected according to the same ID, and the standard triple table data of < entity 1, entity 2 and inter-class relation > is obtained, and data samples are shown in table 1.
If the working voltage of the energy storage cabin No. 1 battery pack No. 1 parallel connection module No. 1 single battery is 3.6V, the dictionary data structure obtained by direct extraction is { 'ID': tank number, 'ID': battery pack number, 'ID': operating voltage, 'ID': the characteristic … is that the corresponding triplets after connection according to the same ID are [ energy storage cabin No. 1- "comprising" -No. 1 single battery ], [ No. 1 single battery- "characteristic" -working voltage 3.6V ], and the like.
In a fifth embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station according to the present invention, based on the fourth embodiment, step S130 further includes the following steps:
step S510: the method comprises the steps of obtaining keywords by adopting manual definition or an automatic learning mode from a corpus of reference texts, and realizing knowledge extraction of a fault maintenance record log by an automatic mode (such as a bootstrap method) or a rule extraction method in a keyword extraction mode.
Specifically, fault types, fault action types, action relations and the like in the data are recorded in a fault maintenance log with keyword prompts, and word frequency is high.
The log data of the fault handling measures is a reference text of fault handling, comprises a large number of common fault types, phenomena and handling modes, is recorded in an official text format, is uniform in content and consistent in editing format, is semi-structured data, and is shown in table 2; table 2 is log data of lithium battery energy storage station fault handling measures.
TABLE 2
Specifically, first, a defined named entity is extracted, and log data (typically, sentences, sentence) is used as input to obtain a corresponding word sequence s=<w 1 ,w 2 ,…,w N >Matching corresponding keywords (e.g. IF using set rules<"xxx+fault">Then<xxx is a fault type entity>、IF<"need+xxx">Then<xxx is action type entity>) Classifying and outputting a set of triples, each tuple being in the form of<I s ,I e ,t>Representing a named entity in s, wherein I s ∈[1,N]And I e ∈[1,N]The start and end positions of the named entity in s are denoted respectively, and t is the entity type. If the input log data is 'battery management system fails', 'needs to be reconnected and a temperature acquisition circuit is tested', the battery management system is obtained through rule matching<1,2,Fault>、<2,3,Action>Namely a "battery management system failed" failure entity type and a "connection and test temperature acquisition circuit" action entity type. And then extracting the corresponding relation types of the two entities, and based on rules such as 'X need Y' (X represents a fault type entity and Y represents an action type entity) 'if the battery management system breaks down and needs to be reconnected and tested by a temperature acquisition circuit', obtaining the corresponding relation type of 'object' by matching rules, and representing the corresponding relation type as 'object' by predicate logic ('the battery management system breaks down', 'the reconnection and the test temperature acquisition circuit').
The above two steps are combined to convert the extracted relationship between entity type and class into triplets to obtain < "faults", "actions", "objects/subjects" >, and the triplets [ the battery management system malfunctions- "objects" -reconnecting and testing the temperature acquisition circuit ] are continuously obtained by the above examples.
In a sixth embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station provided by the invention, based on the fifth embodiment, the expert domain knowledge and rules of fault analysis of the energy storage batteries comprise thesis book related content and expert knowledge text; step S130, further includes the following steps:
step S601: and acquiring the relevant content of the paper book and the expert knowledge text, and cleaning the data to obtain specific fault types, fault phenomenon relevant data parameters, mechanism reason analysis, treatment action modes and fault action schemes of the lithium battery, thereby obtaining a target corpus.
Specifically, the field knowledge and rules of the energy storage battery fault analysis expert are from journal papers and books of lithium battery energy storage station fault research, and access to the knowledge obtained by the field expert/knowledge engineer. The reference data of the fault is the key component for constructing the fault diagnosis knowledge graph of the lithium battery energy storage station. However, the data are explanatory texts, have no fixed word frequency and keyword description, have diversified description forms and are unstructured data.
Step S602: and labeling part of data in the target corpus in an element-label mode.
Specifically, in this embodiment, the labeling mode is BIOES, where B represents the beginning of an entity, I represents the middle of the entity, O represents an irrelevant character, E represents the end of the entity, and S represents a single character.
Step S603: and distinguishing the target corpus by using a specific fault type, fault phenomenon related data parameters, mechanism cause analysis, a treatment action mode and a fault action scheme 5 major entity.
Step S604: randomly dividing the marked partial data into a training set, a testing set and a verification set according to a preset proportion (for example, 8:1:1), and taking the rest unmarked data in the target corpus as a prediction set.
Step S605: the labeled training set is mapped into numerical values through a pre-training language model (such as Skip-gram model in Word2 Vec) to obtain Word vectors.
Step S606: and establishing a loss function according to the sequence tag information to construct an entity tag extraction model.
Specifically, in this embodiment, a commonly used named entity recognition method algorithm framework is constructed through an open source tool TensorFlow, for example, a BiLSTM-CRF algorithm (or an IDCNN-CRF algorithm) is used for realizing entity extraction.
Step S607: the word vector is input into the tag extraction model (i.e., context encoder, such as convolutional neural network CNN) to extract the tags to obtain sequence entity tag information, while the generalization error is estimated by the validation set during training to update the hyper-parameters.
Step S608: and evaluating the constructed entity tag extraction model by using the test set, reconstructing the entity tag extraction model if the evaluation result is lower than a preset target, and terminating the construction of the entity tag extraction model if the evaluation result is greater than or equal to the preset target.
Step S609: and repeatedly changing the structures of the training set, the testing set and the verification set for multiple times, and repeating the steps of constructing the entity tag extraction model to obtain the entity tag extraction model with the best evaluation result.
Step S610: and inputting the rest unlabeled data in the target corpus into an entity tag extraction model to obtain sequence entity tag information, and obtaining corresponding entity types by using a tag decoder (such as a conditional random field CRF decoder).
Specifically, please refer to table 3, table 3 is a sample of knowledge and rules of the energy storage battery fault analysis expert domain.
TABLE 3 Table 3
The data in the table 3 are converted into word vectors through a Skip-gram model and then are input into a trained BiLSTM-CRF entity extraction model to be extracted to obtain the following entities:
fault class: aging of the single batteries, inconsistent aging of the single batteries and aging of the battery pack;
mechanism/cause class: the method comprises the following steps of reducing active lithium ions in a lithium ion battery, increasing SEI film of a negative electrode, forming lithium dendrite (lithium precipitation), decomposing a binder, corroding a current collector, collapsing a structure of a positive electrode, falling active lithium, and generating side reaction between electrolyte and a positive electrode material and corroding the current collector;
action class: reducing the charging multiplying power, controlling the temperature, preventing the overhigh and the overlow, and reducing the cut-off voltage;
class of fault alarm levels: 3 stages;
index class: capacity 50%, failure rate 15%, internal resistance 120mΩ;
finally, determining entity relation according to the predefined relation between classes, for example, the relation between failure class "single battery aging" and mechanism/reason class "reduction of active lithium ions in lithium ion battery" mechanism/reason analysis "is converted into standard triplet [ single battery aging-" mechanism/reason analysis "-reduction of active lithium ions in lithium ion battery ]
In a seventh embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station according to the present invention, based on the sixth embodiment, step S140 includes the following steps:
step S710: and constructing a first knowledge graph based on the triple data corresponding to the structured energy storage station internal historical operation business data.
Step S720: and constructing a second knowledge graph based on the triplet data corresponding to the semi-structured troubleshooting log.
Step S730: and constructing a third knowledge graph based on the unstructured energy storage battery fault analysis expert domain knowledge and the triplet data corresponding to the rule.
Step S740: vectorizing the first knowledge-graph, the second knowledge-graph and the third knowledge-graph through a TransD model.
Specifically, corresponding three-tuple construction is performed to obtain corresponding 3 knowledge maps based on the structured energy storage station internal historical operation service data, the semi-structured fault maintenance log, the unstructured energy storage battery fault analysis expert domain knowledge and the relation between the entities obtained through rule extraction in the sixth embodiment. And then obtaining the vector representation of the triples by the 3 knowledge maps through a TransD model.
In an eighth embodiment of the knowledge graph construction method for fault diagnosis and analysis of the lithium ion battery of the energy storage station, based on the seventh embodiment, the first knowledge graph, the second knowledge graph and the third knowledge graph are all expressed in the form of triples (h, r, t), wherein h represents a head entity, t represents a tail entity and r represents an entity relationship; step S740, including the following steps:
step S801: based on a given correct triplet D + = { (h, r, t) }, initialize entity and relationship embedding.
Step S802: from D + Taking a small set Z of positive facts to generate 2 empty sets
And->
Step S803: for each of ZPositive samples τ + ＝(h + ,r + ,t + ) Generating a negative sample τ according to equation (1) - ＝(h - ,r - ,t - )。
Specifically, the negative sample is obtained by randomly replacing the head entity h or the tail entity t, and E represents an entity set.
Step S804: update set B + ＝B + ∪{τ + Sum B - ＝B - ∪{τ - }。
Step S805: and training a TransD model by using the obtained positive sample and negative sample, regulating parameters by adopting a gradient descent strategy, and embedding updated entities and relations into the { (h, r, t) }.
Step S806: updating the gradient of the loss function and judging whether the gradient is approximately 0.
Step S807: if the gradient is approximately 0, the set Z is selected again, otherwise the output entity and the relation are embedded, wherein the formula (1) is as follows:
the loss function is:
wherein τ= (h, r, t) is D + ∪D - Is a training sample of (a); if (h, r, t) ∈D + Then the following is satisfied: y is hrt =1; if (h, r, t) ∈D - Then the following is satisfied: y is hrt -1; thereby ensuring that the score of the positive sample is higher than that of the negative sample; f (f) r (h, t) is an objective function of the TransD model, and satisfies:
wherein w is r 、w h And w t Mapping vectors additionally introduced by the TransD model, and h is a head entityT is a tail entity, r is an entity relationship, and I is a unit vector.
In a ninth embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station according to the present invention, based on the eighth embodiment, step S807 further includes the following steps:
step S910: and obtaining the representation vectors of the first knowledge graph, the second knowledge graph and the third knowledge graph through the trained TransD model.
Step S920: based on the obtained entity vector, the similarity of the head entity and the tail entity and the similarity of neighbor nodes of the head entity and the tail entity are considered, and the similarity obtained through calculation of the formula (2) is matched with the head entity and the tail entity, so that knowledge map fusion is realized, wherein the formula (2) is as follows:
sim(A,B)＝α×sim Atrr (A,B)+(1-α)×sim NB (A,B)，
wherein A is the representing vector of the head entity, and B is the representing vector of the tail entity; sim (sim) Atrr (A, B) is the similarity of the head entity and the tail entity itself; sim (sim) NB (A, B) is the similarity of neighbor nodes of the head entity and the tail entity.
Step S930: based on the head entity h and the entity relation r, calculating the tail entity t by using a trained TransD model, thereby completing the completion of the knowledge graph.
In a tenth embodiment of the knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries in an energy storage station according to the present invention, based on the first embodiment, the constructed knowledge graph is stored in a Neo4j graph database management system to realize storage and management of the knowledge graph, and the method includes the following steps:
step S1010: and storing the constructed knowledge graph in a Neo4j graph database management system, and updating the database at regular time by using increase and decrease management, informatization searching, structural analysis and decision reasoning.
Specifically, the constructed knowledge graph is stored in the Neo4j graph database management system based on the storage mode of the graph model so as to realize the graph data storage and management functions.
At present, the fault research on the lithium battery energy storage station is limited, and as the energy storage loading is continuously expanded, corresponding fault problems are more and more increased, and the knowledge graph content needs to be updated in a certain time scale. By repeating the steps from S120 to S140, a new knowledge graph can be established, and the functions of content increase and decrease management, informatization searching, structural analysis, decision making and the like of the knowledge graph database can be easily realized by utilizing the advantages of Neo4j that the database is easy to query and edit.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and including several instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present invention and the scope of the claims, which are to be protected by the present invention.
Claims (10)
1. The knowledge graph construction method for the fault diagnosis and analysis of the lithium ion battery of the energy storage station is characterized by comprising the following steps of:
acquiring fault data generated in the operation process of the energy storage station;
extracting fault characteristics of the fault data, thereby defining the relationship between the ontology class and the class of the knowledge graph;
extracting the entity and the relation of the fault data based on the relation between the ontology class and the class to obtain triple data;
constructing a knowledge graph based on the triplet data;
and storing the constructed knowledge graph in a graph database management system to realize the storage and management of the knowledge graph.
2. The knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of an energy storage station according to claim 1, wherein the obtaining fault data generated in the operation process of the energy storage station comprises the following steps:
acquiring structured service data of historical operation in an energy storage station and a semi-structured fault maintenance log, and cleaning the service data and the maintenance log;
and acquiring unstructured energy storage battery fault analysis expert domain knowledge and rules, and cleaning data to obtain specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action schemes.
3. The knowledge graph construction method for diagnosing and analyzing faults of the lithium ion battery of the energy storage station according to claim 2, wherein the extracting fault characteristics of the fault data so as to define ontology classes and relationships among classes of the knowledge graph comprises:
extracting hardware knowledge, fault type knowledge and fault mechanism knowledge of lithium battery equipment according to the acquired fault data, abstracting and defining an ontology class of a knowledge graph, and determining a relation among classes, wherein the ontology class comprises: energy storage cabin class, single battery class, parallel module class, battery pack class, fault class, rule class, action class, value class, voltage class, internal resistance class, internal temperature class, working temperature class, cycle number class, state of charge class, gas class, use duration class, electrolyte internal resistance class and self-discharge rate class; the inter-class relationships include: the method comprises the steps of inclusion, attribute, feature, belongings, subjects or objects, presence or preparation, threshold judgment, alarm level, judgment result, mechanism or reason analysis.
4. A knowledge graph construction method for diagnosing and analyzing faults of lithium ion batteries of an energy storage station according to claim 3, wherein the extracting the entity and the relation of the fault data based on the relation between the body class and the class to obtain the triplet data comprises the following steps:
extracting historical operation business data in the energy storage station to obtain position and operation state information;
and extracting and calculating historical operation service data in the energy storage station to acquire battery operation state data, so as to acquire the information entity of the energy storage equipment and each index type entity, and creating the information entity and each index type entity as a structured knowledge triplet.
5. The knowledge graph construction method for diagnosing and analyzing faults of lithium ion batteries of an energy storage station according to claim 4, wherein the entity and relation of the fault data are extracted based on the relation between the body class and the class to obtain triplet data, and further comprising:
the method comprises the steps of obtaining keywords by adopting manual definition or an automatic learning mode from a corpus of reference texts, and extracting knowledge of a fault maintenance record log in a keyword extraction mode through an automatic mode or a rule extraction method.
6. The knowledge graph construction method for the fault diagnosis and analysis of the lithium ion battery of the energy storage station according to claim 5, wherein the expert domain knowledge and rules of the fault analysis of the energy storage battery comprise thesis book related content and expert knowledge text; the entity and the relation of the fault data are extracted based on the relation between the ontology class and the class so as to obtain triple data, and the method further comprises the following steps:
acquiring the related content of paper books and expert knowledge texts, and cleaning data to obtain specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action schemes of the lithium battery, thereby obtaining a target corpus;
labeling part of data in the target corpus in an element-label mode;
distinguishing the target corpus by using specific fault types, fault phenomenon related data parameters, mechanism cause analysis, treatment action modes and fault action scheme 5 entities;
randomly dividing the marked partial data into a training set, a testing set and a verification set according to a preset proportion, and taking the rest unmarked data in the target corpus as a prediction set;
mapping the marked training set into a numerical value through a pre-training language model to obtain a word vector;
establishing a loss function according to the sequence tag information to construct an entity tag extraction model;
inputting the word vector into the label extraction model so as to extract the label to obtain the sequence entity label information, and estimating the generalization error through the verification set in the training process to update the super-parameters;
evaluating the constructed entity tag extraction model by using the test set, reconstructing the entity tag extraction model if the evaluation result is lower than a preset target, and terminating the construction of the entity tag extraction model if the evaluation result is greater than or equal to the preset target;
and inputting the rest unlabeled data in the target corpus into an entity tag extraction model to obtain sequence entity tag information, and obtaining the corresponding entity type by using a tag decoder.
7. The knowledge graph construction method for diagnosing and analyzing faults of the lithium ion battery of the energy storage station according to claim 6, wherein the knowledge graph construction method based on the triplet data comprises the following steps:
constructing a first knowledge graph based on the triple data corresponding to the structured historical operation business data in the energy storage station;
constructing a second knowledge graph based on the triplet data corresponding to the semi-structured troubleshooting log;
constructing a third knowledge graph based on the unstructured energy storage battery fault analysis expert domain knowledge and the triplet data corresponding to the rule;
vectorizing the first knowledge-graph, the second knowledge-graph and the third knowledge-graph through a TransD model.
8. The knowledge graph construction method for the fault diagnosis and analysis of the lithium ion battery of the energy storage station according to claim 7, wherein the first knowledge graph, the second knowledge graph and the third knowledge graph are all represented in the form of triples (h, r, t), wherein h represents a head entity, t represents a tail entity and r represents an entity relationship; vectorizing the first knowledge-graph, the second knowledge-graph and the third knowledge-graph through a TransD model comprises the following steps:
based on a given correct triplet D + = { (h, r, t) }, initialize entity and relationship embedding;
from D + The set Z of positive facts is obtained to generate 2 empty sets
And->
For each positive sample τ in Z + ＝(h + ,r + ,t + ) Generating a negative sample τ according to equation (1) - ＝(h - ,r - ,t - )；
Update set B + ＝B + ∪{τ + Sum B - ＝B - ∪{τ - }；
Training a TransD model by using the obtained positive sample and negative sample, adjusting parameters by adopting a gradient descent strategy, and embedding an updated entity and relationship into { (h, r, t) };
updating the gradient of the loss function and judging whether the gradient is approximately 0;
if the gradient is approximately 0, the set Z is selected again, otherwise the output entity and the relation are embedded, wherein the formula (1) is as follows:
the loss function is:
wherein τ= (h, r, t) is D + ∪D - Is a training sample of (a); if (h, r, t) ∈D + Then the following is satisfied: y is hrt =1; if (h, r, t) ∈D - Then the following is satisfied: y is hrt -1; thereby ensuring that the score of the positive sample is higher than that of the negative sample; f (f) r (h, t) is an objective function of the TransD model, and satisfies:
wherein w is r 、w h And w t The mapping vectors are additionally introduced by the TransD model, h is a head entity, t is a tail entity, r is an entity relationship, and I is a unit vector.
9. The knowledge graph construction method for diagnosing and analyzing faults of a lithium ion battery of an energy storage station according to claim 8, wherein the step of reselecting the set Z if the gradient is approximately 0, and otherwise, embedding the output entity and the relationship, further comprises:
acquiring representation vectors of the first knowledge graph, the second knowledge graph and the third knowledge graph through a trained TransD model;
based on the obtained entity vector, the similarity of the head entity and the tail entity and the similarity of neighbor nodes of the head entity and the tail entity are considered, and the similarity obtained through calculation of the formula (2) is matched with the head entity and the tail entity, so that knowledge map fusion is realized, wherein the formula (2) is as follows:
sim(A,B)＝α×sim Atrr (A,B)+(1-α)×sim NB (A,B)，
wherein A is the representing vector of the head entity, and B is the representing vector of the tail entity; sim (sim) Atrr (A, B) is the similarity of the head entity and the tail entity itself; sim (sim) NB (a, B) is the similarity of neighbor nodes of the head entity and the tail entity;
based on the head entity h and the entity relation r, calculating the tail entity t by using a trained TransD model, thereby completing the completion of the knowledge graph.
10. The knowledge graph construction method for fault diagnosis and analysis of lithium ion batteries of an energy storage station according to claim 1, wherein the constructed knowledge graph is stored in a Neo4j graph database management system to realize storage and management of the knowledge graph, and the method comprises the following steps:
and storing the constructed knowledge graph in a Neo4j graph database management system, and updating the database at regular time by using increase and decrease management, informatization searching, structural analysis and decision reasoning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310286990.9A CN116384487A (en) | 2023-03-22 | 2023-03-22 | Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310286990.9A CN116384487A (en) | 2023-03-22 | 2023-03-22 | Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116384487A true CN116384487A (en) | 2023-07-04 |
Family
ID=86962722
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310286990.9A Pending CN116384487A (en) | 2023-03-22 | 2023-03-22 | Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116384487A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117077780A (en) * | 2023-10-19 | 2023-11-17 | 国网四川省电力公司信息通信公司 | Evaluation method for optimizing communication private network faults based on particle swarm optimization and knowledge graph |
| CN117113036A (en) * | 2023-10-20 | 2023-11-24 | 中国铁塔股份有限公司吉林省分公司 | Photovoltaic equipment analysis method and system based on energy management system |
| CN117110896A (en) * | 2023-10-24 | 2023-11-24 | 湖北工业大学 | Lithium ion battery fault detection method and system based on knowledge graph |
-
2023
- 2023-03-22 CN CN202310286990.9A patent/CN116384487A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117077780A (en) * | 2023-10-19 | 2023-11-17 | 国网四川省电力公司信息通信公司 | Evaluation method for optimizing communication private network faults based on particle swarm optimization and knowledge graph |
| CN117113036A (en) * | 2023-10-20 | 2023-11-24 | 中国铁塔股份有限公司吉林省分公司 | Photovoltaic equipment analysis method and system based on energy management system |
| CN117113036B (en) * | 2023-10-20 | 2024-01-26 | 中国铁塔股份有限公司吉林省分公司 | Photovoltaic equipment analysis method and system based on energy management system |
| CN117110896A (en) * | 2023-10-24 | 2023-11-24 | 湖北工业大学 | Lithium ion battery fault detection method and system based on knowledge graph |
| CN117110896B (en) * | 2023-10-24 | 2024-01-05 | 湖北工业大学 | Lithium ion battery fault detection method and system based on knowledge graph |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116384487A (en) | Knowledge graph construction method for fault diagnosis and analysis of lithium ion battery of energy storage station | |
| US20220137612A1 (en) | Transformer fault diagnosis and positioning system based on digital twin | |
| Senoussaoui et al. | Combining and comparing various machine‐learning algorithms to improve dissolved gas analysis interpretation | |
| Ha et al. | Technological advances in the fuel cell vehicle: Patent portfolio management | |
| CN110232395A (en) | A kind of fault diagnosis method of electric power system based on failure Chinese text | |
| CN111222330B (en) | Chinese event detection method and system | |
| CN113360582B (en) | Relation classification method and system based on BERT model fusion multi-entity information | |
| CN116108190A (en) | Intelligent operation and maintenance-oriented power transformer knowledge graph construction method | |
| CN116310551A (en) | Power transformer fault diagnosis method and system of multi-mode information fusion network | |
| Shi et al. | Fault diagnosis of signal equipment on the lanzhou-xinjiang high-speed railway using machine learning for natural language processing | |
| Savelka et al. | Cross-domain generalization and knowledge transfer in transformers trained on legal data | |
| Tian et al. | Method for predicting the remaining mileage of electric vehicles based on dimension expansion and model fusion | |
| Goldberg et al. | CASTLE: crowd-assisted system for text labeling and extraction | |
| Jose et al. | Finding best answer in community question answering sites: a review | |
| CN117216617A (en) | Text classification model training method, device, computer equipment and storage medium | |
| Wen et al. | A novel auto-lstm-based state of health estimation method for lithium-ion batteries | |
| Wan et al. | Evaluation model of power operation and maintenance based on text emotion analysis | |
| CN115619117A (en) | Power grid intelligent scheduling method based on duty system | |
| CN114117078A (en) | Electric power heterogeneous knowledge fusion method based on D-S evidence theory | |
| Li et al. | A new identify disruptive technologies algorithm based on technology develop network | |
| Chen et al. | State-of-charge estimation of lithium-ion batteries using convolutional neural network with self-attention mechanism | |
| Leung et al. | The Application of Affective Measures in Text-based Emotion Aware Recommender Systems | |
| Chai et al. | Towards conversational qa: automatic identification of problematic situations and user intent | |
| Hantono et al. | LSTM for State of Charge Estimation of Lithium Polymer Battery on Jetson Nano | |
| CN117110896B (en) | Lithium ion battery fault detection method and system based on knowledge graph |
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 |