CN116340534A - Knowledge graph construction method and system for identifying new energy abnormal data - Google Patents
Knowledge graph construction method and system for identifying new energy abnormal data Download PDFInfo
- Publication number
- CN116340534A CN116340534A CN202310198191.6A CN202310198191A CN116340534A CN 116340534 A CN116340534 A CN 116340534A CN 202310198191 A CN202310198191 A CN 202310198191A CN 116340534 A CN116340534 A CN 116340534A
- Authority
- CN
- China
- Prior art keywords
- data
- knowledge graph
- abnormal
- new energy
- relationship
- 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
Images
Classifications
-
- 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
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Computational Linguistics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Probability & Statistics with Applications (AREA)
- Algebra (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a knowledge graph construction method and a knowledge graph construction system for identifying new energy abnormal data. The method comprises the following steps: the knowledge graph is utilized to represent the abnormal data information and the related relation thereof, the limitation that the traditional method is difficult to adapt to the abnormal condition of complex new energy data is broken through, the information required for constructing the abnormal data knowledge graph is extracted on the basis of fully combing the internal logic of the abnormal data, so that the construction of the knowledge graph is realized, the method for automatically tracking the abnormal data is provided, and the overall process management of 'abnormal data occurrence-abnormal positioning-abnormal judgment-abnormal tracing' based on the knowledge graph is realized.
Description
Technical Field
The invention belongs to the field of new energy and knowledge graph data preprocessing, and particularly relates to a knowledge graph construction method and system for identifying new energy abnormal data.
Background
The production data collected on site of the new energy station not only can directly reflect the running condition of production equipment, but also is the basis for carrying out data analysis, fault diagnosis, power prediction, production management and other works. However, in the production process of the new energy station, due to the reasons of complex geographical environment, changeable weather conditions, numerous production conditions and the like, such as abnormal communication, extreme weather, equipment faults, human factors and the like, the quality of the acquired production data of a plurality of energy stations is poor, the actual production conditions of the stations can be seriously deviated by directly utilizing the abnormal data, the actual operation conditions of the new energy station equipment can not be accurately and effectively reflected, and the operation management and the scheduling of the new energy station are adversely affected; in addition, because the abnormal data cannot be timely distinguished and the fault reasons cannot be traced back, the real rule between the abnormal data and the generated reasons is difficult to find, so that an effective method for solving the abnormal data cannot be found, and the trouble is half caused by tasting.
Disclosure of Invention
In order to solve the technical problems, the invention provides a technical scheme of a knowledge graph construction method for identifying new energy abnormal data, so as to solve the technical problems.
The first aspect of the invention discloses a knowledge graph construction method for identifying new energy abnormal data, which comprises the following steps:
step S1, an overall architecture of a new energy abnormal data knowledge graph, namely G= (E, R, S, U), is formed by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
s2, constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
s3, constructing an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data;
and step S4, correlating the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
According to the method of the first aspect of the present invention, in the step S2, the object domain knowledge graph is composed of a plurality of triples, and represents "entity-relationship-entity", "entity-relationship-attribute" or "concept-attribute value";
The entity comprises: power stations and equipment;
the concept includes: region, power station type, equipment type, season, month and weather type;
the attributes include: capacity, longitude and latitude, grid-connected time, grid-connected voltage level and rainfall in manufacturers, models and weather types of equipment of a power station and enterprises to which the power station belongs;
the attribute value refers to the value of the object domain entity object or the concept related attribute.
According to the method of the first aspect of the present invention, in the step S2, the method for constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source includes:
and carrying out knowledge extraction on the triples: extracting entities and attributes of a power station, equipment, a measuring point position and a geographic position of the new energy source; extracting the relation between the entity and the attribute of the new energy power station;
carrying out knowledge fusion on the extracted entity, attribute and relationship between the entity and the attribute;
and (3) performing rechecking processing on the indirect entity relationship of the entity, the attribute and the relationship after knowledge fusion, namely, providing the indirect entity relationship to obtain an object domain knowledge graph.
According to the method of the first aspect of the present invention, in the step S2, the relationship between the entity and the attribute of the power station includes:
Causal relationship, which refers to the logical relationship of results caused by reasons;
a dependency relationship, which refers to a relationship between a component and a device;
inclusion relationships, meaning conceptual and phenomenological inclusion relationships;
action relationship, meaning some action that the device plays;
parallel relationship refers to the parallel relationship between equipment and components;
precedence relationship, which refers to the precedence order between devices;
isotactic relation, meaning that different phrases represent the same word sense;
positional relationship, which represents a relationship in position;
cross condition, meaning that when a device belongs to one sequence, it will not belong to another sequence;
other relationships cannot be expressed directly by the above relationship.
According to the method of the first aspect of the present invention, in the step S3, the abnormal event domain knowledge graph includes: unique identification of the abnormal data event, data type, data threshold, timestamp of occurrence of the abnormal data, specific measuring point position where the abnormal data occurs and abnormal data type description.
According to the method of the first aspect of the present invention, in the step S3, the method for constructing the abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data includes:
Establishing an abnormal data type base according to the time sequence data curve, the log file and the meteorological data;
in the abnormal data type library, setting a threshold value of data by adopting a predefined method according to the data item and the data type;
according to the KKS code of the power station, a unique mark is assigned to each measuring point of the abnormal data, namely, a unique identifier of an abnormal data event;
and selecting a predefined algorithm according to the data item and the data type, judging the abnormal data according to a set threshold value, selecting the abnormal data, and generating an abnormal data set, namely an abnormal event domain knowledge graph.
According to the method of the first aspect of the present invention, in the step S3, the abnormal data type library includes a data item, a data type, a data threshold value, and a data abnormal type; the data items comprise generation power, generation energy, current value and voltage value; the data types include integer type, floating point type, character type and Boolean type; the data exception types include null values, error values, negative values, transition values, and extremum values.
The invention discloses a knowledge graph construction system for identifying new energy abnormal data, which comprises the following steps:
the first processing module is configured to form a general framework of a new energy abnormal data knowledge graph, namely G= (E, R, S, U) by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
The second processing module is configured to construct the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
the third processing module is configured to construct an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data;
and the fourth processing module is configured to correlate the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
According to the system of the second aspect of the present invention, the second processing module is configured to make the object domain knowledge graph consist of a plurality of triples, and represent "entity-relationship-entity", "entity-relationship-attribute" or "concept-attribute value";
the entity comprises: power stations and equipment;
the concept includes: region, power station type, equipment type, season, month and weather type;
the attributes include: capacity, longitude and latitude, grid-connected time, grid-connected voltage level and rainfall in manufacturers, models and weather types of equipment of a power station and enterprises to which the power station belongs;
the attribute value refers to the value of the object domain entity object or the concept related attribute.
According to the system of the second aspect of the present invention, the second processing module is configured to construct the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy, and the method includes:
and carrying out knowledge extraction on the triples: extracting entities and attributes of a power station, equipment, a measuring point position and a geographic position of the new energy source; extracting the relation between the entity and the attribute of the new energy power station;
carrying out knowledge fusion on the extracted entity, attribute and relationship between the entity and the attribute;
and (3) performing rechecking processing on the indirect entity relationship of the entity, the attribute and the relationship after knowledge fusion, namely, providing the indirect entity relationship to obtain an object domain knowledge graph.
According to the system of the second aspect of the present invention, the second processing module is configured to, in relation to the entity and the attribute of the power station, include:
causal relationship, which refers to the logical relationship of results caused by reasons;
a dependency relationship, which refers to a relationship between a component and a device;
inclusion relationships, meaning conceptual and phenomenological inclusion relationships;
action relationship, meaning some action that the device plays;
parallel relationship refers to the parallel relationship between equipment and components;
Precedence relationship, which refers to the precedence order between devices;
isotactic relation, meaning that different phrases represent the same word sense;
positional relationship, which represents a relationship in position;
cross condition, meaning that when a device belongs to one sequence, it will not belong to another sequence;
other relationships cannot be expressed directly by the above relationship.
According to the system of the second aspect of the present invention, the third processing module is configured to include: unique identification of the abnormal data event, data type, data threshold, timestamp of occurrence of the abnormal data, specific measuring point position where the abnormal data occurs and abnormal data type description.
According to the system of the second aspect of the present invention, the third processing module is configured to construct, from bottom up, an abnormal event domain knowledge graph according to the time series data curve, the log file and the meteorological data, and includes:
establishing an abnormal data type base according to the time sequence data curve, the log file and the meteorological data;
in the abnormal data type library, setting a threshold value of data by adopting a predefined method according to the data item and the data type;
according to the KKS code of the power station, a unique mark is assigned to each measuring point of the abnormal data, namely, a unique identifier of an abnormal data event;
And selecting a predefined algorithm according to the data item and the data type, judging the abnormal data according to a set threshold value, selecting the abnormal data, and generating an abnormal data set, namely an abnormal event domain knowledge graph.
A third aspect of the invention discloses an electronic device. The electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps in the knowledge graph construction method for identifying the new energy source abnormal data in any one of the first aspects of the disclosure when executing the computer program.
A fourth aspect of the invention discloses a computer-readable storage medium. A computer-readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps in a knowledge graph construction method of identifying new energy anomaly data of any one of the first aspects of the present disclosure.
The scheme provided by the invention can be used for guiding new energy power station owners, operation and maintenance personnel, scientific researchers and the like to conveniently establish the correlation between station equipment, measuring point positions, geographic environments, weather conditions and the like and abnormal data, is convenient for searching the generation reasons of the abnormal data, reduces equipment fault discrimination time, improves the operation and maintenance efficiency of the power station, and realizes the quality improvement and synergy of the power station.
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 needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a knowledge graph construction method for identifying new energy anomaly data according to an embodiment of the present invention;
FIG. 2 is a block diagram of a knowledge graph construction system for identifying new energy anomaly data according to an embodiment of the present invention;
fig. 3 is a block diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The first aspect of the invention discloses a knowledge graph construction method for identifying new energy abnormal data. Fig. 1 is a flowchart of a knowledge graph construction method for identifying new energy abnormal data according to an embodiment of the present invention, as shown in fig. 1, the method includes:
step S1, an overall architecture of a new energy abnormal data knowledge graph, namely G= (E, R, S, U), is formed by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
wherein, the g= (E, R, S) triplet is a general representation of the knowledge graph. E is an entity set, R is a relation set,
basic form of S: (entity 1, relationship, entity 2), (entity, attribute value), and the like.
S2, constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
s3, constructing an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data;
and step S4, correlating the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
In step S1, an overall architecture of a new energy abnormal data knowledge graph, namely g= (E, R, S, U), is formed by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data.
Specifically, the knowledge graph is composed of various entities and relations of the new energy station, and has corresponding data abnormal conditions and types, and the object domain knowledge graph and the abnormal event domain knowledge graph form the overall architecture of the new energy abnormal data knowledge graph, namely G= (E, R, S, U), which represents related factors such as power stations, equipment, measuring points, attributes, formed reasons and the like related to abnormal data; the object domain knowledge graph consists of triples, namely G= (E, R, S), wherein E represents an entity set, R represents a relation set, S represents a triplet set, and the triples are related marks, codes, attributes and association relations of parts such as new energy stations, equipment, measuring point positions, geographic positions and the like; u is an abnormal event domain knowledge graph, and represents the type and condition of abnormal data, namely the type of abnormal data related to the abnormal data part, and the factors such as date, position, reason and the like and the causal relationship related to each other are generated.
And in step S2, constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy.
In some embodiments, in the step S2, the object domain knowledge graph is composed of a plurality of triples, which represent "entity-relationship-entity", "entity-relationship-attribute" or "concept-attribute value";
the entity comprises: power stations and equipment;
the concept includes: region, power station type, equipment type, season, month and weather type;
the attributes include: capacity, longitude and latitude, grid-connected time, grid-connected voltage level and rainfall in manufacturers, models and weather types of equipment of a power station and enterprises to which the power station belongs;
the attribute value refers to the value of the object domain entity object or the concept related attribute.
The method for constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy comprises the following steps:
and carrying out knowledge extraction on the triples: extracting entities and attributes of a power station, equipment, a measuring point position and a geographic position of the new energy source; extracting the relation between the entity and the attribute of the new energy power station;
Carrying out knowledge fusion on the extracted entity, attribute and relationship between the entity and the attribute;
and (3) performing rechecking processing on the indirect entity relationship of the entity, the attribute and the relationship after knowledge fusion, namely, providing the indirect entity relationship to obtain an object domain knowledge graph.
The relationship between the entity and the attribute of the power station comprises:
causal relationship, which refers to the logical relationship of results caused by reasons;
a dependency relationship, which refers to a relationship between a component and a device;
inclusion relationships, meaning conceptual and phenomenological inclusion relationships;
action relationship, meaning some action that the device plays;
parallel relationship refers to the parallel relationship between equipment and components;
precedence relationship, which refers to the precedence order between devices;
isotactic relation, meaning that different phrases represent the same word sense;
positional relationship, which represents a relationship in position;
cross condition, meaning that when a device belongs to one sequence, it will not belong to another sequence;
other relationships cannot be expressed directly by the above relationship.
Specifically, the object domain knowledge graph is composed of triplets, that is, g= (E, R, S) described above. Wherein E= { E 1 ,e 2 ,…,e n -representing a collection of entities in the knowledge graph of the object domain, n entities in total, each entity being identifiable with a uniquely determined ID; r= { R 1 ,r 2 ,…,r m The method comprises the steps of representing a set of various semantic relations in an object domain knowledge graph, wherein m relations exist in a form of directed edges connecting entity head nodes and entity tail nodes;
representative objectAnd collecting in the part domain knowledge graph. Wherein E is head Head node representing entity, E tail Representing the tail node of the entity.
The object domain knowledge graph consists of a plurality of triples, and represents an entity-relationship-entity, an entity-relationship-attribute or a concept-attribute value and the like. The entity is the most basic constituent element in the object domain knowledge graph, and comprises a power station, equipment (such as a string, an inverter, a combiner box and the like in a photovoltaic power station) and the like; concepts mainly include sets, categories, types, categories, etc., such as territories (province, city, county), power station types (centralized, distributed), inverter types (centralized, cluster), seasons (spring, summer, autumn, winter), months (1-12 months), weather types (sunny, cloudy, rain, snow, etc.), etc.; the attributes mainly refer to characteristics, features, parameters and the like related to physical objects or concepts of the object domain, power stations (capacity, longitude and latitude, grid-connected time, grid-connected voltage level, enterprises) and equipment (manufacturer and model), rainy days (light rain, medium rain, heavy rain, rain and snow), and the like; the attribute values mainly refer to the values of the physical objects of the object domain or the related attributes of the concept, such as the capacity of a power station of "6MW", the grid-connected voltage level of "10kV", the rated output power of an inverter of "500kW", the month of the genus "7 months", and the daily irradiance of about "5.85kWh/m 2 "etc. In this way, each triplet is associated with each other to form an object domain knowledge graph containing knowledge. The specific method comprises the following steps:
and (5) carrying out knowledge extraction. The method mainly extracts entities, attributes and relations contained in unstructured and semi-structured new energy power station data, marks parts of speech and uses the parts of speech as basic elements for constructing a new energy object domain knowledge graph. Knowledge extraction is divided into entity extraction, relation extraction and attribute extraction.
Firstly, extracting the entity and attribute related to the new energy power station.
(1) And performing word segmentation. By means of data such as a common word dictionary, an electric power dictionary, new energy station information, equipment information, weather information, power station operation records, power station operation and maintenance records and the like, a hidden Markov model lambda= (A, B, pi) and a corpus to be segmented are utilizedColumn o= (O) 1 ,o 2 ,…,o x ) And obtaining accurate word segmentation. Wherein:
a represents a state transition probability matrix between parts of speech,
a ij indicating that the part of speech is at p at time t i State, t+1 is part of speech p j Probability of state. i=1, 2, …, k; j=1, 2, …, k. Each part of speech corresponds to a state, and the transition probabilities of the states represent the relationships between the parts of speech.
B represents the part-of-speech to word observation probability matrix, 
Indicating that the part of speech is at p at time t i The probability of the corresponding word q is generated in the state, which is also called the generation probability. q=1, 2, …, l; j=1, 2, …, k.
Pi represents the prior probability of part of speech, i.e. part of speech initial state probability vector, pi= (pi) i ) Indicating that at time t=1, the part of speech p is present i Probability of state.
o x Representing X sentences in the corpus to be segmented.
Thus, when a word is segmented for a certain unstructured and semi-structured information in the language, the state is continuously transferred through the hidden Markov model, and each state can generate an output, namely a phrase or a word, until the output of the whole information is completed.
By analyzing all the corpus, the occurrence frequency of each part of speech, the phrase or word corresponding to the part of speech, each part of speech and the occurrence frequency of the subsequent part of speech are obtained. Converting the frequency into probability, obtaining the sum of each probability by calculating the probability of word occurrence and the connection probability of part of speech, and selecting the probability with the largest probability as an output result.
(2) Extracting phrase. And performing professional retrieval on the phrase or word obtained after the word segmentation by the hidden Markov model in a general dictionary or a special power dictionary, and extracting the phrase or word as an entity, concept or attribute of the object domain knowledge graph if the response content can be retrieved.
(3) Part of speech identification. At the same time of phrase extraction, the searched phrases or words are labeled in part of speech, such as nouns (for example, an inverter is used as a noun and is marked by a noun), verbs (for example, an inverter rated alternating current output power is less than 500kW and is marked by a verb) representing attributes, adverbs (for example, a photovoltaic power station daily generating capacity is often lower than an average value and is marked by an adv), prepositions (for example, a photovoltaic power station is located in an east city area of Beijing and is marked by prep) and conjunctions (for example, a power station generating capacity is low because of shielding and is marked by conj).
And extracting the relation of the related entity and attribute of the new energy power station.
The relation extraction is mainly used for judging whether the relation exists between each entity and attribute in the new energy power station and the corresponding relation type. The relation between the entities and the attributes related to the new energy power station is divided into 10 relations, and the relation mainly comprises the following steps:
(1) Causality, which is a logical relationship of results caused by reasons, such as "impact" which can lead to "hidden cracks";
(2) A subordinate relationship refers to a relationship between components and devices, for example, a "DC/AC converter" belongs to an "inverter";
(3) Inclusion relationships, meaning conceptual, phenomenologically inclusion relationships, such as "hot spots", "hidden cracks" are all "faults";
(4) Functional relationship, meaning that the device performs some function, such as "combiner box" for "collecting current";
(5) Parallel relationship refers to parallel relationship between equipment and components, such as "A term current", "B term current", "C term current" in alternating current;
(6) The precedence relationship refers to the precedence order among devices, for example, "string inverter" before "combiner box" and "centralized inverter" after "combiner box".
(7) Isotactic relation means that different phrases represent the same meaning, such as "radiant flux" also known as "radiant power".
(8) Positional relationship indicates positional relationship, for example, "photovoltaic module" is located on "stand".
(9) Cross-condition, meaning that when a device belongs to one sequence, it does not belong to another sequence, such as a "fixed stent" and a "tracking stent"; the tracking bracket can be divided into a single-shaft type bracket and a double-shaft type bracket;
(10) Other relationships cannot be expressed directly by the above relationship.
These relationships, in combination with entities, concepts and attributes, are represented by a number of triplet base units (E head ，R，E tail ) And the part domain knowledge maps are connected to form the embryonic form of the part domain knowledge maps.
And carrying out knowledge fusion. The knowledge extraction of the new energy object domain knowledge graph is completed to obtain information such as entity, concept, relationship and attribute from unstructured and semi-structured data of new energy, but many redundant or erroneous information may exist in the results, so that knowledge fusion is required.
The entity disambiguates. In the process of constructing the new energy object domain knowledge graph, the problem that a certain entity name item corresponds to a plurality of named entity objects exists, for example, the entity name of sunlight can correspond to sunlight, company name, power station name and inverter manufacturer name. By means of entity disambiguation, entity names corresponding to the phrases can be accurately established according to the context in which the words are located. We mainly employ entity disambiguation methods based on entity links here.
The input content of the entity disambiguation method based on entity link mainly comprises two parts, namely an object domain knowledge base obtained through training in the steps, and mainly comprises the following steps: entity, concept description, entity attributes, context semantic information, and the like. And secondly, the entity item to be disambiguated and the context related information thereof.
First establishing a candidate set E of link entity reference items y ,E y ＝{e 1 ,e 2 ,…,e x }，y≤n，
Selecting the entity items f and E to be disambiguated y Establishing a relation, calculating E by using argmax function y The entity with the largest score in the relevance to f is taken as the target entity of the entity item to be disambiguated. Namely:
The Score (e, f) is calculated by using the topic consistency model, and is mainly used for judging the consistency degree of the entity item to be disambiguated and other entities in the knowledge base. First, the importance degree of each entity in the knowledge base context, namely the relevance of the entity to the semantic topic, is calculated. The average value of semantic association of a certain entity and other entities is calculated and is mainly used as the weight of the importance degree of the entity.
Wherein T refers to an entity, T refers to a set of all entities in the context of each entity, and β (e, T) refers to the weight of entity e; sr (e, e) x ) Refers to entity e and entity e x Semantic association values between.
Then, the consistency degree of the entity item to be disambiguated and other entities of the context is calculated. The weighted semantic association average value of the entity item to be disambiguated and other entities is mainly calculated to be used as the consistency score.
Coreference resolution. Co-resolution is mainly used to solve the case where multiple entity reference items correspond to the same entity. Such as "solar cell" and "photovoltaic cell" refer to the same thing, as well as cell arrays, also known as photovoltaic arrays or cell arrays, and the like.
Part of speech classification is first performed. Since the parts of speech of the plural entity reference items are necessarily the same, we can divide all the words representing the entity and the attribute into Z sets according to different parts of speech, and respectively perform synonym recognition on each set.
In order to show the semantic similarity of the plurality of entity referring items, a Skip-Gram model of a word2vec algorithm is adopted, and the word vector dimension is set to 300 dimensions, so that word vectors corresponding to all words in different sets are obtained.
And judging the similarity degree among the plurality of entity index items by calculating the cosine similarity among the word vectors.
Where g represents a g-dimensional space, C, D is two g-dimensional vectors, C is [ C1, C2, ], cg ], D is [ D1, D2, ], dg ]. C·d represents the dot product (vector product) of two vectors, |c|×|d| represents the product of two vector lengths (or sizes).
Through calculation, the position adjacent words in the context or the context parity words have high cosine similarity. However, the co-word is the synonym we want to determine, therefore, neighboring words are culled and the co-word is preserved.
Forming a synonym pair. And merging and sorting the screened co-located words containing the same word sense to form a synonym pair, and representing by adopting a standardized name.
And performing rechecking processing of indirect entity relationship on the entity, attribute and relationship after knowledge fusion. And eliminating some redundant attributes and redundant relations. For example, according to the above steps, direct entity relationship triplets of entities { inverter, fan, failure }, { fan, slave, inverter }, and { [ fan, occurrence, failure ] } are generated, but indirect entity relationship triplets of { inverter, occurrence, failure } may also be formed; to reduce the complexity of knowledge-graph, the triplet forms of these indirect relationships are deleted, and if necessary, can be inferred and identified by direct entity relationships.
And in step S3, constructing an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data.
In some embodiments, in the step S3, the abnormal event domain knowledge graph includes: unique identification of the abnormal data event, data type, data threshold, timestamp of occurrence of the abnormal data, specific measuring point position where the abnormal data occurs and abnormal data type description.
The method for constructing the abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data comprises the following steps:
Establishing an abnormal data type base according to the time sequence data curve, the log file and the meteorological data;
in the abnormal data type library, setting a threshold value of data by adopting a predefined method according to the data item and the data type;
according to the KKS code of the power station, a unique mark is assigned to each measuring point of the abnormal data, namely, a unique identifier of an abnormal data event;
and selecting a predefined algorithm according to the data item and the data type, judging the abnormal data according to a set threshold value, selecting the abnormal data, and generating an abnormal data set, namely an abnormal event domain knowledge graph.
The abnormal data type library comprises data items, data types, data thresholds and data abnormal types; the data items comprise generation power, generation energy, current value and voltage value; the data types include integer type, floating point type, character type and Boolean type; the data exception types include null values, error values, negative values, transition values, and extremum values.
Specifically, the event domain knowledge graph includes a unique identifier of an abnormal data event, a data type, a data threshold, a timestamp of occurrence of abnormal data, a specific measurement point position where the abnormal data occurs, an abnormal data type description, and the like.
Firstly, an abnormal data type base is established, and mainly comprises data items, data types, data thresholds and data abnormal types. For example, the data items include power generation power, power generation amount, current value, voltage value, etc., the data types include integer type, floating point type, character type, boolean type, etc., and the data anomaly types include null value, error value, negative value, jump value, extremum, etc., for example, "daily power generation amount exceeds theoretical power generation amount value", "daytime power generation data is negative value", etc.
According to different data items and data types, different data thresholds are set by different methods, for example, a statistical quartile range (interquartile range, IQR) method can be used to calculate first and third quartiles (Q1, Q3), and abnormal data is outside Q1, Q3, i.e.
Num ab ＞Q3+K(IQR)∨Num ab ＜Q1-K(IQR)
Wherein Num is ab Representing anomalous data, iqr=q3-q1k+.0, for example, here we can prefer k=1.5
A Z-score method may also be used to determine if anomalies by calculating how far the data is from the average. The data threshold is
μ is the average of all relevant data, σ is the standard deviation of all relevant data. After normalization, if |num ab And if the I is larger than Zthr, the data is considered to be abnormal. Here we can choose a Zthr value of 2.5 in a limited way.
In addition, we can also judge abnormal data by Principal Component Analysis (PCA), K-means clustering algorithm, LOF (local outlier factor) algorithm, etc.
According to the KKS code of the power station, a unique mark code is assigned to each measuring point, and meanwhile, the mark code can also be used as a unique mark code of a data source.
Selecting a certain time period, selecting different algorithms according to different thresholds for different data items and data typesJudging row abnormal data, selecting abnormal data, generating an abnormal data set, wherein U= { U 1 ,u 2 ,…,u v }，v∈N。
And step S4, correlating the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
Specifically, for abnormal data generated in a certain time period, a candidate relation set of the object domain entity and the abnormal data event is generated. First, according to the abnormal data set u= { U 1 ,u 2 ,…,u v V E N, distinguishing through a unique mark code, and selecting an entity set of a measuring point related to abnormal data, wherein E= { E 1 ,e 2 ,…,e n }，n∈N。
And matching the abnormal data with the entities of the related measuring points one by one. There are cases where one measurement point entity corresponds to one abnormal data value, or one measurement point entity corresponds to a plurality of abnormal data values.
A set of relationships of the object domain entity to the abnormal data event is generated. According to different measuring points, different abnormal data sets U' = { U are generated 1 ',u' 2 ,…,u' v }，
By judging the relation established in the step 02-1-2, a set with a correlation relation is selected, and a corresponding relation set R' = { R is obtained 1 ',r 2 ',…,r v '}，/>
A relationship topology of the anomalous data events for the individual measurement points is generated. The relation set topology G '= (E', R ', U') of a single measuring point is established, so that the information of the types and the abnormal values of the abnormal data generated on a certain measuring point, the information of the equipment and the parts related to each abnormal data, the information of the equipment attribute, the external environment, the weather condition and the like corresponding to the abnormal data and the like can be searched, and a reference can be provided for judging the generation reason of the abnormal data.
Generating object domain-event domain knowledge maps corresponding to all abnormal data: and connecting and fusing all the measuring points generating the abnormal data to obtain an object domain directed map of the abnormal data, and connecting measuring point entities of the object domains with the abnormal data event. And obtaining object domain-event domain knowledge graphs of all abnormal data in a certain time period, namely, the total knowledge graph G= (E, R, S, U) of the abnormal data of the new energy.
In summary, the scheme provided by the invention can be used for guiding new energy power station owners, operation and maintenance personnel, scientific researchers and the like to conveniently establish the correlation between station equipment, measuring point positions, geographic environments, weather conditions and the like and abnormal data, is convenient for searching for reasons for generating the abnormal data, reduces equipment fault discrimination time, improves the operation and maintenance efficiency of the power station, and realizes quality improvement and synergy of the power station.
The invention discloses a knowledge graph construction system for identifying new energy abnormal data. FIG. 2 is a block diagram of a knowledge graph construction system for identifying new energy anomaly data according to an embodiment of the present invention; as shown in fig. 2, the system 100 includes:
a first processing module 101 configured to form an overall architecture of a new energy anomaly data knowledge graph, i.e., g= (E, R, S, U) from the object domain knowledge graph and the anomaly event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
the second processing module 102 is configured to construct the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
A third processing module 103 configured to construct an abnormal event domain knowledge graph from bottom to top according to the time series data curve, the log file and the meteorological data;
the fourth processing module 104 is configured to correlate the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph g= (E, R, S, U).
According to the system of the second aspect of the present invention, the second processing module 102 is configured to make the object domain knowledge graph comprise a plurality of triples, which represent "entity-relationship-entity", "entity-relationship-attribute" or "concept-attribute value";
the entity comprises: power stations and equipment;
the concept includes: region, power station type, equipment type, season, month and weather type;
the attributes include: capacity, longitude and latitude, grid-connected time, grid-connected voltage level and rainfall in manufacturers, models and weather types of equipment of a power station and enterprises to which the power station belongs;
the attribute value refers to the value of the object domain entity object or the concept related attribute.
According to the system of the second aspect of the present invention, the second processing module 102 is configured to construct the object domain knowledge graph from top to bottom according to the power station, the device, the measuring point position and the geographic position of the new energy, where the constructing includes:
And carrying out knowledge extraction on the triples: extracting entities and attributes of a power station, equipment, a measuring point position and a geographic position of the new energy source; extracting the relation between the entity and the attribute of the new energy power station;
carrying out knowledge fusion on the extracted entity, attribute and relationship between the entity and the attribute;
and (3) performing rechecking processing on the indirect entity relationship of the entity, the attribute and the relationship after knowledge fusion, namely, providing the indirect entity relationship to obtain an object domain knowledge graph.
According to the system of the second aspect of the present invention, the second processing module 102 is configured to, in relation to the entity and the attribute of the power station, include:
causal relationship, which refers to the logical relationship of results caused by reasons;
a dependency relationship, which refers to a relationship between a component and a device;
inclusion relationships, meaning conceptual and phenomenological inclusion relationships;
action relationship, meaning some action that the device plays;
parallel relationship refers to the parallel relationship between equipment and components;
precedence relationship, which refers to the precedence order between devices;
isotactic relation, meaning that different phrases represent the same word sense;
positional relationship, which represents a relationship in position;
cross condition, meaning that when a device belongs to one sequence, it will not belong to another sequence;
Other relationships cannot be expressed directly by the above relationship.
According to the system of the second aspect of the present invention, the third processing module 103 is configured to include: unique identification of the abnormal data event, data type, data threshold, timestamp of occurrence of the abnormal data, specific measuring point position where the abnormal data occurs and abnormal data type description.
According to the system of the second aspect of the present invention, the third processing module 103 is configured to construct, from bottom up, an abnormal event domain knowledge graph according to the time series data curve, the log file and the meteorological data, where the constructing includes:
establishing an abnormal data type base according to the time sequence data curve, the log file and the meteorological data;
in the abnormal data type library, setting a threshold value of data by adopting a predefined method according to the data item and the data type;
according to the KKS code of the power station, a unique mark is assigned to each measuring point of the abnormal data, namely, a unique identifier of an abnormal data event;
and selecting a predefined algorithm according to the data item and the data type, judging the abnormal data according to a set threshold value, selecting the abnormal data, and generating an abnormal data set, namely an abnormal event domain knowledge graph.
A third aspect of the invention discloses an electronic device. The electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the steps in the knowledge graph construction method for identifying the new energy abnormal data according to any one of the first aspect of the disclosure when executing the computer program.
Fig. 3 is a block diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 3, the electronic device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the electronic device is used for conducting wired or wireless communication with an external terminal, and the wireless communication can be achieved through WIFI, an operator network, near Field Communication (NFC) or other technologies. The display screen of the electronic equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the electronic equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the structure shown in fig. 3 is merely a structural diagram of a portion related to the technical solution of the present disclosure, and does not constitute a limitation of the electronic device to which the present application is applied, and that a specific electronic device may include more or less components than those shown in the drawings, or may combine some components, or have different component arrangements.
A fourth aspect of the invention discloses a computer-readable storage medium. The computer readable storage medium stores a computer program which, when executed by a processor, implements the steps in a knowledge graph construction method for identifying new energy anomaly data according to any one of the first aspect of the present disclosure.
Note that the technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be regarded as the scope of the description. The foregoing examples represent only a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (10)
1. The method for constructing the knowledge graph for identifying the abnormal data of the new energy is characterized by comprising the following steps of:
step S1, an overall architecture of a new energy abnormal data knowledge graph, namely G= (E, R, S, U), is formed by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
s2, constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
s3, constructing an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data;
and step S4, correlating the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
2. The knowledge graph construction method for recognizing new energy abnormal data according to claim 1, wherein in the step S2, the object domain knowledge graph is composed of a plurality of triples, and represents "entity-relationship-entity", "entity-relationship-attribute" or "concept-attribute value";
The entity comprises: power stations and equipment;
the concept includes: region, power station type, equipment type, season, month and weather type;
the attributes include: capacity, longitude and latitude, grid-connected time, grid-connected voltage level and rainfall in manufacturers, models and weather types of equipment of a power station and enterprises to which the power station belongs;
the attribute value refers to the value of the object domain entity object or the concept related attribute.
3. The knowledge graph construction method for recognizing new energy anomaly data according to claim 2, wherein in the step S2, the method for constructing the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy comprises:
and carrying out knowledge extraction on the triples: extracting entities and attributes of a power station, equipment, a measuring point position and a geographic position of the new energy source; extracting the relation between the entity and the attribute of the new energy power station;
carrying out knowledge fusion on the extracted entity, attribute and relationship between the entity and the attribute;
and (3) performing rechecking processing on the indirect entity relationship of the entity, the attribute and the relationship after knowledge fusion, namely, providing the indirect entity relationship to obtain an object domain knowledge graph.
4. A knowledge graph construction method for identifying new energy anomaly data according to claim 3, wherein in step S2, the relationship between the entity and the attribute of the power station comprises:
causal relationship, which refers to the logical relationship of results caused by reasons;
a dependency relationship, which refers to a relationship between a component and a device;
inclusion relationships, meaning conceptual and phenomenological inclusion relationships;
action relationship, meaning some action that the device plays;
parallel relationship refers to the parallel relationship between equipment and components;
precedence relationship, which refers to the precedence order between devices;
isotactic relation, meaning that different phrases represent the same word sense;
positional relationship, which represents a relationship in position;
cross condition, meaning that when a device belongs to one sequence, it will not belong to another sequence;
other relationships cannot be expressed directly by the above relationship.
5. The knowledge graph construction method for recognizing new energy anomaly data according to claim 1, wherein in the step S3, the anomaly event domain knowledge graph comprises: unique identification of the abnormal data event, data type, data threshold, timestamp of occurrence of the abnormal data, specific measuring point position where the abnormal data occurs and abnormal data type description.
6. The method for constructing a knowledge graph for identifying abnormal data of new energy according to claim 5, wherein in the step S3, the method for constructing the knowledge graph of the abnormal event domain from bottom to top according to the time series data curve, the log file and the meteorological data comprises:
establishing an abnormal data type base according to the time sequence data curve, the log file and the meteorological data;
in the abnormal data type library, setting a threshold value of data by adopting a predefined method according to the data item and the data type;
according to the KKS code of the power station, a unique mark is assigned to each measuring point of the abnormal data, namely, a unique identifier of an abnormal data event;
and selecting a predefined algorithm according to the data item and the data type, judging the abnormal data according to a set threshold value, selecting the abnormal data, and generating an abnormal data set, namely an abnormal event domain knowledge graph.
7. The knowledge graph construction method for recognizing new energy anomaly data according to claim 6, wherein in the step S3, the anomaly data type library includes data items, data types, data thresholds, and data anomaly types; the data items comprise generation power, generation energy, current value and voltage value; the data types include integer type, floating point type, character type and Boolean type; the data exception types include null values, error values, negative values, transition values, and extremum values.
8. A knowledge graph construction system for identifying new energy anomaly data, the system comprising:
the first processing module is configured to form a general framework of a new energy abnormal data knowledge graph, namely G= (E, R, S, U) by an object domain knowledge graph and an abnormal event domain knowledge graph; the object domain knowledge graph consists of triples, namely G= (E, R, S); u is an abnormal event domain knowledge graph and represents the type and condition of abnormal data;
the second processing module is configured to construct the object domain knowledge graph from top to bottom according to the power station, the equipment, the measuring point position and the geographic position of the new energy source;
the third processing module is configured to construct an abnormal event domain knowledge graph from bottom to top according to the time sequence data curve, the log file and the meteorological data;
and the fourth processing module is configured to correlate the object domain knowledge graph with the abnormal event domain knowledge graph to obtain a new energy abnormal data overall knowledge graph G= (E, R, S, U).
9. An electronic device comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps in a knowledge graph construction method of identifying new energy anomaly data according to any one of claims 1 to 7 when the computer program is executed.
10. A computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and the computer program, when executed by a processor, implements the steps in a knowledge-graph construction method for identifying new energy anomaly data according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310198191.6A CN116340534A (en) | 2023-02-27 | 2023-02-27 | Knowledge graph construction method and system for identifying new energy abnormal data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310198191.6A CN116340534A (en) | 2023-02-27 | 2023-02-27 | Knowledge graph construction method and system for identifying new energy abnormal data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116340534A true CN116340534A (en) | 2023-06-27 |
Family
ID=86892237
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310198191.6A Pending CN116340534A (en) | 2023-02-27 | 2023-02-27 | Knowledge graph construction method and system for identifying new energy abnormal data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116340534A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116521904A (en) * | 2023-06-29 | 2023-08-01 | 湖南大学 | Ship manufacturing data cloud fusion method and system based on 5G edge calculation |
-
2023
- 2023-02-27 CN CN202310198191.6A patent/CN116340534A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116521904A (en) * | 2023-06-29 | 2023-08-01 | 湖南大学 | Ship manufacturing data cloud fusion method and system based on 5G edge calculation |
| CN116521904B (en) * | 2023-06-29 | 2023-09-22 | 湖南大学 | Ship manufacturing data cloud fusion method and system based on 5G edge calculation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP2024502730A (en) | Medical data element automated classification method and system based on depth map matching | |
| CN116245107B (en) | Electric power audit text entity identification method, device, equipment and storage medium | |
| CN113869052A (en) | AI-based house address matching method, storage medium and equipment | |
| CN114358014A (en) | Work order intelligent diagnosis method, device, equipment and medium based on natural language | |
| CN116340534A (en) | Knowledge graph construction method and system for identifying new energy abnormal data | |
| CN113505190B (en) | Address information correction method, device, computer equipment and storage medium | |
| CN114491081A (en) | Electric power data tracing method and system based on data blood relationship graph | |
| CN114155044A (en) | Power price prediction method and system for power spot market node | |
| CN118069839A (en) | Cross-language event detection method based on graph neural network | |
| CN117290404A (en) | Method and system for rapidly searching and practical main distribution network fault processing method | |
| CN117010373A (en) | Recommendation method for category and group to which asset management data of power equipment belong | |
| CN113988083B (en) | Factual information coding and evaluating method for generating shipping news abstract | |
| Hu et al. | A classification model of power operation inspection defect texts based on graph convolutional network | |
| CN111814457B (en) | Power grid engineering contract text generation method | |
| CN115409122A (en) | Method, system, equipment and medium for analyzing concurrent faults of power transformation equipment | |
| Zhang et al. | Robust graph learning under Wasserstein uncertainty | |
| CN111339258B (en) | University computer basic exercise recommendation method based on knowledge graph | |
| Gong | Analysis of internet public opinion popularity trend based on a deep neural network | |
| Li et al. | Multi-source heterogeneous log fusion technology of power information system based on big data and imprecise reasoning theory | |
| Zheng | Individualized Recommendation Method of Multimedia Network Teaching Resources Based on Classification Algorithm in a Smart University | |
| CN117272053B (en) | Method for generating address data set with few samples, address matching method, medium and equipment | |
| Zhang et al. | A multivariate time series anomaly detection model based on graph attention mechanism in energy consumption of intelligent buildings | |
| Cao et al. | Research on deep mining of MOOC multimodal resources based on improved Eclat algorithm | |
| Chen et al. | Public demand estimation following disasters through integrating social media and community demographics | |
| Yaşar et al. | Estimating spatiotemporal focus of documents using entropy with PMI |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication |