CN113792538B - Method and device for rapidly generating operation ticket of power distribution network - Google Patents
Method and device for rapidly generating operation ticket of power distribution network Download PDFInfo
- Publication number
- CN113792538B CN113792538B CN202110994717.2A CN202110994717A CN113792538B CN 113792538 B CN113792538 B CN 113792538B CN 202110994717 A CN202110994717 A CN 202110994717A CN 113792538 B CN113792538 B CN 113792538B
- Authority
- CN
- China
- Prior art keywords
- model
- state
- distribution network
- environment
- state matrix
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000009826 distribution Methods 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 44
- 239000011159 matrix material Substances 0.000 claims abstract description 84
- 230000011218 segmentation Effects 0.000 claims abstract description 48
- 238000013135 deep learning Methods 0.000 claims abstract description 38
- 238000012423 maintenance Methods 0.000 claims abstract description 34
- 238000004458 analytical method Methods 0.000 claims abstract description 31
- 238000012549 training Methods 0.000 claims description 34
- 230000005540 biological transmission Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 9
- 230000002457 bidirectional effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000002372 labelling Methods 0.000 abstract description 12
- 238000010586 diagram Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- 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
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Water Supply & Treatment (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method and a device for rapidly generating operation tickets of a power distribution network. The method comprises the following steps: sequentially inputting a maintenance application form sentence text into a word segmentation model, a part-of-speech labeling model, a dependency syntax analysis model, a distribution network scheduling semantic role recognition model and a maintenance application form semantic recognition model, and outputting key information of operation tasks in different types of maintenance application forms; and converting the initial state and the termination state of the equipment in the key information into an environment initial state matrix and an environment termination state matrix, and then inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model for step-by-step reasoning to obtain all ordered operation items corresponding to the operation task. The invention can intelligently generate the dispatching instruction ticket, converts the artificial writing into automatic ticket forming of the system, releases the labor capacity of distribution network regulation personnel, and greatly saves ticket forming time.
Description
Technical Field
The invention relates to a method and a device for rapidly generating operation tickets of a power distribution network, and belongs to the field of regulation and control management for the power distribution network.
Background
The operation ticket is an operation instruction set aiming at specific equipment and specific operation tasks, and related equipment running states are switched according to the power grid safety operation rules and the description sequence of the operation terms.
When a dispatcher opens a dispatching operation ticket, firstly opening an overhaul application form, checking the corresponding equipment name and the running state, and checking the content of the overhaul application form; then analyzing the content of the overhaul application form, and extracting an operation task; searching a historical typical operation ticket in a dispatching management system according to an operation task, opening a related circuit diagram to analyze and check the running state of a circuit, and carrying out manual reasoning operation items according to the corresponding equipment name and topological relation and the historical typical ticket experience rule; and finally, manually filling the operation items item by item to generate an operation ticket.
Along with the increasing of the distribution network scale, the control and management difficulty is also increased continuously. The distribution network is directly connected with the vast power customers, how to improve the power supply reliability, and the 'no power failure and little power failure' of thousands of households becomes the first thing considered by distribution network regulation management staff. The current wiring mode of the power distribution network is more and more complex, the scheduling operation ticket is formulated more and more frequently, a scheduler is more and more difficult to control the increasingly complex power distribution network system, and the problem of safe and stable operation of the power distribution network is more and more serious. At present, the operation ticket rule customization is complex, the existing operation ticket management system cannot realize online real-time billing, and strict verification is lacking in the billing process. The billing work occupies a large amount of time for regulating and controlling the service, the time for generating the overhaul operation ticket is required to be shortened, the task amount of dispatching staff is reduced, and a new operation ticket generating method is provided.
Disclosure of Invention
The invention aims to provide a rapid generation method of an operation ticket of a power distribution network, which aims to solve the problems of complex operation and long ticket forming time in the manual ticket forming process in the prior art.
In order to solve the technical problems, the invention adopts the following technical scheme:
In one aspect, the invention provides a method for rapidly generating an operation ticket of a power distribution network, which comprises the following steps:
Inputting the work content, the type of the overhaul ticket and the sentence text in the stop or power transmission range rail in the overhaul application form into a conditional random field word segmentation model, and outputting a word list after word segmentation;
Inputting the word list subjected to word segmentation into a conditional random field part-of-speech tagging model, and outputting a word list tagged with part-of-speech;
Inputting the word list marked with the part of speech into a dependency syntax analysis model, and outputting a word list containing the dependency relationship of each word;
inputting the word list containing the word dependency relationship into a distribution network scheduling semantic role identification model, and outputting a word list marked with distribution network scheduling semantic role attributes;
Inputting the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic identification model, and outputting key information of operation tasks in different types of maintenance application forms;
converting the equipment starting state and the termination state in the key information into an environment initial state matrix and an environment termination state matrix;
and inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to perform step-by-step reasoning, so as to obtain all ordered operation items corresponding to the operation task.
Further, the conditional random field word segmentation model is trained by using a quasi-Newton algorithm, and the corpus used for training comprises: beijing university word segmentation corpus, ICWB conference word segmentation corpus, MSR05 word segmentation corpus and related corpus for power distribution network scheduling of an electric power system.
Further, the conditional random field part-of-speech tagging model is trained by a Viterbi algorithm, the part-of-speech tagging standard is a standard of word segmentation and part-of-speech tagging of modern Chinese corpus processing Specification of Beijing university, and the part-of-speech is divided by a coarse granularity standard.
Further, the dependency syntactic analysis model is trained by a neural network algorithm, the corpus used for training comprises a semantic dependency network corpus tree library of Qinghai university and related corpora scheduled by a power system power distribution network, each sentence is marked according to CONLL format, and marking content comprises sequence numbers, words, prototypes, coarse-granularity parts of speech, fine-granularity parts of speech, syntactic features, center words and dependency relations.
Further, the distribution network scheduling semantic role identification model comprises a bidirectional long-short time memory network and a conditional random field layer connected with the bidirectional long-short time memory network, and the semantic role attribute labeling method comprises the following steps: and carrying out one-hot coding according to the positions of the words in the distribution network scheduling initial dictionary, and taking the coding as the label of the distribution network scheduling semantic roles.
Further, inputting the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model, and outputting key information of operation tasks in different types of maintenance application forms, wherein the method specifically comprises the following steps:
In a word list marked with distribution network scheduling semantic role attributes, performing paraphrasing search on the content of the operation type to determine the type of the overhaul application form;
Extracting station names, feeder line names, equipment names related to operation and equipment state names from a word list marked with distribution network scheduling semantic role attributes according to the type of the overhaul application form; searching a mediator relation in the dependency relation, and searching for a device name of a starting position and a terminating position of a power outage section or an operation section, a device name of a new device start/exit and a state name before and after the device operation from the device names related to the operation; and searching for parallel relations in the dependency relation, and dividing sentences containing the parallel relations into a plurality of short sentences to obtain key information of the operation task of the overhaul application form.
Further, the converting the initial state and the termination state of the device in the key information into an environment initial state matrix and an environment termination state matrix specifically includes:
And calling a ring network topological structure according to the equipment ID in the key information, and combining the equipment starting state and the termination state in the key information with the ring network topological structure to obtain an environment initial state matrix and an environment termination state matrix.
Further, the step-by-step reasoning is performed by inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model, so as to obtain all the ordered operation items corresponding to the operation task, which specifically comprises the following steps:
Inputting the environment initial state matrix into a deep learning reasoning network model, outputting a first operation item state by the model, and comparing the input environment initial state matrix with the output change of the first operation item state to obtain a first operation item; continuously inputting the first operation item state into a deep learning reasoning network model, outputting to obtain a second operation item state, and comparing the input first operation item state with the output second operation item state change to obtain a second operation item; and analogically, until the operation item state output by the model is the same as the environment termination state matrix, and obtaining all ordered operation items corresponding to the operation task.
Further, the foregoing method for rapidly generating the operation ticket of the power distribution network further includes, after each operation item is obtained: and (3) carrying out error-proof verification on the obtained operation items, if the verification is successful, continuously inputting the operation item final state output by the model into a deep learning reasoning network model to output the operation item final state of the next step, and if the verification is unsuccessful, recovering the operation item final state of the previous step and inputting the operation item final state into the deep learning reasoning network model again to carry out reasoning.
Further, the deep learning reasoning network model is trained by adopting a deep deterministic strategy gradient algorithm, and the training method of the deep learning reasoning network model comprises the following steps:
Acquiring a plurality of historical operation tickets;
For each history operation ticket, calling a ring network topological structure according to the contained equipment ID;
Based on a ring network topological structure, converting the initial state and the termination state of the equipment corresponding to each historical operation ticket into an environment initial state matrix X 0 and an environment termination state matrix X e;
And inputting an environment initial state matrix X 0 corresponding to each historical operation ticket and states X 1、X2、…、XT of all operation items, wherein T is the number of the operation items, respectively inputting a deep learning reasoning network model, sequentially outputting a variable 1 and a variable 2 … variable T by the model, and training the deep learning reasoning network model by taking the states that the variable 1 is the same as the state X 1 of the first operation item, the variable 2 is the same as the state X 2 of the second operation item, … and the variable T is the same as the state matrix X e of the environment termination.
Further, the method for converting the environment state matrix comprises the following steps:
For a single feeder tree network, assuming that it contains n single-ended devices and m double-ended devices, the states of the single-ended devices are represented by an n×1 matrix S:
S＝[x1 x2 x3…xn]T
Wherein: x 1 represents the state of the first single-ended device, the values 0,1, 2 or 3,0 representing "running", 1 representing "hot standby", 2 representing "cold standby", 3 representing "service";
the state of the double-ended device is represented by a matrix D of 1×m:
D＝[e1 e2 e3…em]
Wherein e 1 represents the state of the first two-terminal device, the values 0,1, 2 or 3,0 representing "running", 1 representing "hot standby", 2 representing "cold standby", 3 representing "maintenance";
according to the ring network topological structure, the matrix S and the matrix D are associated to obtain an environment state matrix K:
Wherein: the row represents a single-ended device, the column represents a double-ended device, k 1,1 represents the relationship between the first double-ended device and the first single-ended device, and k 1,1 takes on the values 1, 0, -1, where 1 represents that the first single-ended device is the head end of the first double-ended device, -1 represents that the first single-ended device is the end of the first double-ended device, and 0 represents no relationship.
In another aspect, the present invention provides a rapid generation device for an operation ticket of a power distribution network, including:
The word segmentation module is configured to input the work content, the type of the overhaul ticket and the sentence text in the stop or power transmission range rail in the overhaul application form into the conditional random field word segmentation model and output a word list after word segmentation;
the part-of-speech tagging module is configured to input the word list subjected to word segmentation into a conditional random field part-of-speech tagging model and output a word list tagged with part-of-speech;
The dependency syntax analysis module is configured to input the word list marked with the part of speech into a dependency syntax analysis model and output a word list containing the dependency relation of each word;
The semantic role recognition module is configured to input the word list containing the word dependency relationship into a distribution network scheduling semantic role recognition model and output a word list marked with distribution network scheduling semantic role attributes;
The semantic recognition module is configured to input the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model and output key information of operation tasks in different types of maintenance application forms;
the preprocessing module is configured to convert the equipment starting state and the equipment ending state in the key information into an environment initial state matrix and an environment ending state matrix;
and the reasoning module is configured to input the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to perform step-by-step reasoning so as to obtain all ordered operation items corresponding to the operation task.
The invention achieves the beneficial technical effects that: (1) Carrying out semantic analysis on texts of the overhaul application form and the task term by adopting a natural language processing technology, releasing the labor capacity of network allocation regulation personnel, and simultaneously realizing the association of the operation ticket and the application form; (2) The scheduling instruction ticket is intelligently generated by combining the graph data, the artificial sketch is converted into the automatic ticket forming of the system, and the working time of the sketch is greatly saved.
Drawings
Fig. 1 is a schematic flow chart of a method for rapidly generating operation tickets of a power distribution network in an embodiment of the invention;
FIG. 2 is an exemplary diagram of a service application form;
Fig. 3 is a diagram illustrating a topology.
Detailed Description
The invention is further described below in connection with specific embodiments. The following examples are only for more clearly illustrating the technical aspects of the present invention, and are not intended to limit the scope of the present invention.
As described above, when a dispatcher issues a dispatch ticket, a manual operation is required to generate the ticket, but the manual ticket generation operation is complicated and the ticket generation time is long.
To this end, an embodiment of the present invention provides a method for rapidly generating an operation ticket of a power distribution network, as shown in fig. 1, where the method includes the following steps:
step 11, inputting the work content, the overhaul ticket type and the sentence text in the stop or power transmission range rail in the overhaul application form into a conditional random field word segmentation model, and outputting a word list after word segmentation; firstly, training a conditional random field word segmentation model by adopting a quasi-Newton algorithm.
The word segmentation model uses a generic corpus comprising: "Beijing university word segmentation corpus", "ICWB conference word segmentation corpus", "MSR05 (Microsoft research institute) word segmentation corpus".
In addition, related corpora of power system distribution network scheduling, such as overhaul contents in historical overhaul application forms of test point units and contents of distribution network regulation and control operation manuals and regulations of local power companies, are collected and used as corpora for training word segmentation models.
And carrying out Chinese word segmentation labeling on the general corpus and the related corpus scheduled by the power system power distribution network by using a BMSE four-structure word position labeling method, wherein B represents the beginning of a word, E represents the end of the word, M represents the middle of the word, and S represents the word formed independently.
Inputting the marked general corpus and the related corpus scheduled by the power system power distribution network into a word segmentation model for training, and storing a model file after training.
And then, when the trained word segmentation model is used for word segmentation, inputting the text of the maintenance application statement into the model, and outputting a word list after word segmentation.
Step 12, inputting the word list subjected to word segmentation into a conditional random field part-of-speech tagging model, and outputting a word list tagged with part-of-speech;
Firstly, training a conditional random field part-of-speech tagging model by adopting a Viterbi algorithm.
The part-of-speech labeling is consistent with the part-of-speech content of the word segmentation part, and comprises general corpus and power system distribution network dispatching related corpus, and the part-of-speech labeling is added on the basis of the word segmentation corpus.
The standard of part of speech tagging adopts the standard of word segmentation and part of speech tagging of modern Chinese corpus processing specification of Beijing university, and adopts the standard of coarse granularity to divide the part of speech.
And inputting the labeled corpus into a part-of-speech labeling model for training, and storing a model file after training.
And then, when the trained part-of-speech tagging model is used for part-of-speech tagging, inputting the sentences of the overhaul application form after word segmentation into the part-of-speech tagging model, and outputting a word list with part of speech.
Step 13, inputting the word list marked with the part of speech into a dependency syntax analysis model, and outputting a word list containing the dependency relationship of each word;
First, a neural network algorithm is used to train a dependency syntax analysis model.
The general corpus of the dependency syntactic analysis is a semantic dependency network corpus tree library of the Qinghai university. And collecting historical overhaul application forms of test point units and statement contents in power system distribution network dispatching related materials.
Each sentence in the corpus is labeled according to CONLL format, the labeling content comprises serial number, word, prototype (Chinese word=prototype), coarse-granularity part of speech, fine-granularity part of speech, syntactic feature, central word (serial number of central word in sentence) and dependency relationship, because the part of speech of coarse granularity has been labeled in the part of speech labeling training process, the part of speech of coarse granularity and the part of speech of fine granularity in CONLL format are consistent with the part of speech of the corpus in the part of speech labeling training process.
Inputting the annotated general dependency syntax corpus and the dependency syntax corpus related to power system distribution network scheduling into a dependency syntax analysis model for training, and storing a model file after training.
After training, the dependency syntax analysis can be performed by using the dependency syntax analysis model, the word list with part of speech in the sentence is input into the dependency syntax analysis model, the dependency relation of each word in the sentence is obtained, and the output format is CONLL format.
Step 14, inputting the word list containing the word dependency relationship into a distribution network scheduling semantic role identification model, and outputting a word list marked with distribution network scheduling semantic role attributes;
The distribution network scheduling semantic character recognition model adopts BiLSTM-CRF model, and comprises a bidirectional long and short time memory network (BiLSTM) layer and a Conditional Random Field (CRF) layer connected with the bidirectional long and short time memory network. The BiLSTM layer predicts the label score of each word unit semantic role, and the CRF layer adds sequence constraint to the predicted labels so as to improve the accuracy of model identification.
First, biLSTM-CRF models are trained.
Based on the corpus of the CONLL format dependency syntactic analysis of the step 13, adding labels of the single semantic roles of the overhaul application of each word. Wherein, the semantic roles include: station name, feeder name, equipment name related to operation, equipment status name, operation type (e.g. power outage, power transmission), etc.
In a specific embodiment, one-hot coding is performed according to the position of each word in the scheduling initial dictionary, and the one-hot coding is used as a label of a scheduling semantic role of the distribution network.
Inputting the corpus after adding the labels into BiLSTM-CRF model for training, and storing model files after training.
After training, a BiLSTM-CRF model can be used for carrying out distribution network dispatching semantic role recognition, sentences in CONLL format are input, the BiLSTM layer predicts label scores of semantic roles of each word, all scores predicted by the BiLSTM layer are input into the CRF layer, and the highest label score in all semantic roles of each word unit is used as the semantic role of the word in the CRF layer.
Step 15, inputting the word list marked with the distribution network dispatching semantic role attribute into a maintenance application form semantic recognition model, and outputting key information of operation tasks in different types of maintenance application forms;
after the maintenance application form sentences are subjected to word segmentation, part-of-speech tagging, dependency syntactic analysis and distribution network scheduling semantic role recognition, key information is recognized according to different maintenance application form categories based on the analysis output results, and finally, maintenance application form semantic analysis is completed.
Firstly, performing a near meaning word search on the contents in an operation type column in an overhaul application form, dividing the overhaul application form into four types of power failure type, power transmission type, mode adjustment type and new equipment start/exit type, and extracting key information from four types of overhaul application form sentences according to the analysis result:
a) The key information that the service request form of the power failure class needs to be extracted includes: station name, feeder line name, operation type, power failure interval start position equipment name, power failure interval end position equipment name, pre-operation start state and post-operation end state.
B) The key information to be extracted from the power transmission type overhaul application form comprises: station name, feeder line name, operation type, power transmission section start position equipment name, power transmission section end position equipment name, pre-operation start state, and post-operation end state.
C) The key information to be extracted for the mode adjustment type overhaul application form comprises the following steps: station name, feeder line name, mode adjustment type, operation interval starting position equipment name, operation interval ending position equipment name, pre-operation starting state and post-operation ending state.
D) The key information required to be extracted by the new equipment start/exit type overhaul application form comprises the following steps: station name, feeder name, start/exit device name, operation type.
After the steps 11-14, each maintenance application form sentence is divided into words, each word contains the results of part-of-speech tagging, dependency syntactic analysis and maintenance application form semantic role tagging, and sentences containing the results are used as input for the semantic recognition of the maintenance application form.
Thus, this step may specifically comprise:
Step 151, performing a near-meaning word search on the operation type in the result of the semantic role labeling, and determining the type of the overhaul application form;
And 152, extracting key information of an operation task of the overhaul application form according to the type of the overhaul application form.
Firstly, extracting information such as station names, feeder line names, equipment names related to operation, equipment state names and the like from semantic role labeling results;
then searching for a mediator relation in the dependency relation, and searching for a device name of a starting position and a terminating position of a power outage section or an operation section, a device name of a new device start/exit and a state name before and after the device operation in the device names related to the operation;
And finally, searching for parallel relations in the dependency relation, and dividing sentences containing the parallel relations into a plurality of short sentences to finish the extraction of the key information of the maintenance application form sentences.
When the overhaul application form rule of the test point unit changes, the key information extraction rule can be modified according to actual conditions.
The final output of the maintenance application form semantic analysis is a key information list in the containing statement, such as { "station name": du Gang station, "feeder name": sentry line, "operation type": overhauling, namely an equipment name of an initial position of an operation section: #47 rod drop-out fuse, "operation section termination position device name": end line, "pre-operation start state": running, "post-operation termination state": overhaul }.
In addition, the information except for the 'work content' in the overhaul application form is in a structured table form, and can be directly extracted according to the actual condition of the test point unit, for example: the order-setting type, the operation purpose, the time of construction, the time of completion, the service ticket application ID and the like.
And step 16, converting the equipment starting state and the termination state in the key information into an environment initial state matrix and an environment termination state matrix.
In one embodiment, a ring network topology structure is called according to the device ID in the key information, and the device start state and the device end state in the key information are combined with the ring network topology structure to obtain an environment initial state matrix and an environment end state matrix. For an electrical network, the electrical devices, single-ended devices (switches) and double-ended devices (lines and transformers), are connected into a wiring diagram, and each device has a different state. Each of the operation items in the operation ticket corresponds to a change in the state of the electrical device.
The electrical operating code specifies that the electrical device has four basic operating states; operation, hot standby, cold standby and maintenance status:
(1) Operating state: the circuit breaker corresponding to the equipment and the disconnecting link at the two sides of the circuit breaker are in a closing state, and in the closing state, the equipment has certain voltage;
(2) Hot standby state: the breaker of the equipment is positioned at the opening position, and the knife switches at the two sides are in the closing state, so long as the breaker is closed, the equipment is immediately changed into the running state;
(3) Cold standby state: the breaker and the knife switches at both sides are positioned at the opening position;
(4) And (3) maintenance: the breaker and the two side disconnecting link of the equipment are required to be positioned at the disconnecting link position, and the grounding disconnecting link is required to be closed.
On the basis, preprocessing is carried out on the data, and the history operation ticket and the topological structure information are processed into an environment state matrix of the deep learning reasoning network model.
Specifically, in a single feeder tree network, the data preprocessing includes the following processes:
(1) Single-ended device-switch
Let n be the total number of switches in the topology, the switch state is represented by a matrix S of n x 1:
S＝[x1 x2 x3…xn]T
Wherein: x 1 represents the state of the first switch, and may take the values 0,1, 2 or 3,0 representing "running", 1 representing "hot standby", 2 representing "cold standby", and 3 representing "service".
(2) Double-ended device-line and transformer
Let the total number of double-ended devices be m, the double-ended device state is represented by a matrix D of 1×m:
D＝[e1 e2 e3…em]
Wherein e 1 represents the state of the first two-terminal device, the values 0, 1, 2 or 3,0 representing "running", 1 representing "hot standby", 2 representing "cold standby", 3 representing "maintenance"; when the line considers whether FA is put in, 0 represents "run+fa put in", 1 represents "hot standby", 2 represents "cold standby", 3 represents "overhaul", and 4 represents "run+fa inactive";
(3) Topological relation
The topological relation is represented by an n x m directed correlation matrix K:
Wherein: the row represents a switch, the column represents a line or transformer, k 1,1 represents the relationship between the first line or transformer and the first switch, k 1,1 takes the values 1, 0, -1, where 1 represents the switch is the head end of the line or transformer, -1 represents the switch is the end of the line or transformer, and 0 represents no association.
For example, as shown in fig. 3, the topology matrix thereof may be expressed as:
(4) Environment state matrix X
The environment state matrix X of the deep learning inference network model simultaneously contains the state and topological relation of the equipment, and is expressed as follows:
and step 17, inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to perform step-by-step reasoning, so as to obtain all ordered operation items corresponding to the operation task.
In the step, the deep learning reasoning network model adopts a depth deterministic strategy Gradient algorithm (DEEP DETERMINISTIC Policy Gradient, DDPG) to intelligently infer operation items.
The intelligent operation item reasoning is to use the operation ticket rules as the basis, and to obtain the ordered operation items corresponding to the operation tasks by reasoning according to the operation task key information and the power grid topological structure obtained by semantic analysis.
Training of the model is required prior to reasoning using DDPG deep learning reasoning network models.
Before model training, firstly, data preprocessing is required to be carried out on the historical operation ticket, the historical operation ticket and topological structure information are processed into an environment state matrix of DDPG deep learning reasoning network model, the environment state matrix is used as input of model training, and the specific processing process is as described in step 16.
Training of DDPG deep learning reasoning network model is performed by taking each operation ticket as a training round, and the process is as follows:
① Inputting a history operation ticket, and calling a topology analysis service according to the equipment ID to acquire a ring network topology structure;
② And combining the ring network topological structure, and converting the initial state and the termination state of the electrical equipment in the operation task to obtain an environment state initial matrix X 0 and a termination state matrix X e.
③ If the operation items in the operation ticket share T items, the state of the first operation item is X 1 and …, the state of the T-th operation item is X T, and X 0、X1、…、XT、Xe is sequentially input into the reasoning model to complete one round of training.
The operation item is changed from X 0 to X 1, from X 1 to X 2……,XT-1 to X T; the operational item state refers to the final state of the operational item (i.e., change), i.e., X 1、…、XT.
After training with a large number of historical operation tickets, the model file is saved.
And after training, reasoning key information in the operation task obtained through semantic analysis in the step 15 by using DDPG deep learning reasoning network model, so as to generate an operation ticket containing the ordered operation items. The specific flow is as follows:
① The key information in the operation task obtained by the semantic analysis of the overhaul application form comprises equipment ID, equipment starting and ending states and the like; and calling a topology analysis service according to the device ID, acquiring a ring network topology structure, and processing the starting state and the ending state of the device to obtain an environment state initial matrix X 0 and an ending state matrix X e.
② Inputting X 0 and X e into a DDPG deep learning reasoning network model, and carrying out model reasoning step by step.
③ The reasoning is further carried out, an intermediate variable X t is output, the content of the operation item can be obtained by comparing the change A t of X t-1 and X t, if a certain switch state is changed from 0 to 1, the operation item is corresponding to: "change a switch from run to hot standby". ";
④ Comparing X t with termination state X e, if X t＝Xe, the reasoning ends, if X t and X e do not agree, return to ③.
In a further embodiment, an "anti-error check" service may be invoked to perform "five-prevention" check on the operation item a t obtained by inference, if the check is successful, continuing the next inference, if the check is unsuccessful, recovering the state X t-1, and returning to the ③ step for re-inference.
In addition, in the reasoning process, specified operation items can be added in combination with the special requirements of the trial unit ground city.
According to the power distribution network operation ticket rapid generation method, the content of the overhaul application ticket is analyzed through the semantic recognition technology, key information such as the overhaul power failure range, the states before and after operation, the operation types and the like is extracted, the operation items are generated based on intelligent reasoning of the expert knowledge base, the expert knowledge base rules are formed through combing, the operation items are automatically generated for the operation tasks extracted from the overhaul application ticket through training to generate the reasoning model, the operation is simple, the labor force is saved, and the ticket forming time is shortened.
On the basis of the method, the operation process can be simulated and previewed through a graphical human-computer interface.
The method comprises the following steps:
based on the graphical human-computer interface, the operation is automatically performed sequentially according to the content of the operation ticket, the state of the operation object is automatically changed, and whether each operation item is correct or not is automatically checked through simulation operation.
And initializing the replay line state according to the initial state of the operation ticket replay.
The system automatically loads the operation instructions of the operation ticket, performs simulation previewing one by one, and can truly reflect the actual process executed by the operation ticket by changing the state of the operation object; and provide the following functions of breakpoint setting, single-step previewing, automatic previewing, recovering, etc.:
automatic previewing: the system automatically executes all operation ticket instructions in the operation ticket according to the operation sequence, and displays the state of each operation object after the execution is finished through the human-computer interface.
Setting a breakpoint: and setting power failure in one operation instruction in the operation ticket, and automatically executing all instructions before the breakpoint according to the operation sequence by the system if the breakpoint exists during automatic previewing and displaying the state of each current operation object on a human-computer interface.
Single step previewing: and on the basis of the previous operation instruction, sequentially executing the previous operation instruction, and displaying the states of all operation objects after execution on a human-computer interface.
And (5) recovering: and restoring the state of each operation object before the execution of the operation instruction in the operation ticket.
After the replay is finished, the replay can be repeated, the initial state is requested to be recovered, the replay line is initialized, and then the simulation replay is carried out piece by piece.
The operation ticket based on the graph simulates the previewing, the contents in the operation ticket can be simulated and executed in the graph in sequence, the running condition of the power grid after the operation is intuitively displayed, the misjudgment and misexecution conditions of a dispatcher are reduced, and the dispatching safety running level is effectively improved.
In another embodiment, the present invention provides a rapid generating device for an operation ticket of a power distribution network, including:
The word segmentation module is configured to input the work content, the type of the overhaul ticket and the sentence text in the stop or power transmission range rail in the overhaul application form into the conditional random field word segmentation model and output a word list after word segmentation;
the part-of-speech tagging module is configured to input the word list subjected to word segmentation into a conditional random field part-of-speech tagging model and output a word list tagged with part-of-speech;
The dependency syntax analysis module is configured to input the word list marked with the part of speech into a dependency syntax analysis model and output a word list containing the dependency relation of each word;
The semantic role recognition module is configured to input the word list containing the word dependency relationship into a distribution network scheduling semantic role recognition model and output a word list marked with distribution network scheduling semantic role attributes;
The semantic recognition module is configured to input the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model and output key information of operation tasks in different types of maintenance application forms;
the preprocessing module is configured to convert the equipment starting state and the equipment ending state in the key information into an environment initial state matrix and an environment ending state matrix;
and the reasoning module is configured to input the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to perform step-by-step reasoning so as to obtain all ordered operation items corresponding to the operation task.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is merely a preferred embodiment of the present invention, and it should be noted that modifications and variations could be made by those skilled in the art without departing from the technical principles of the present invention, and such modifications and variations should also be regarded as being within the scope of the invention.
Claims (7)
1. The rapid generation method of the operation ticket of the power distribution network is characterized by comprising the following steps of:
Inputting the work content, the type of the overhaul ticket and the sentence text in the stop or power transmission range rail in the overhaul application form into a conditional random field word segmentation model, and outputting a word list after word segmentation;
Inputting the word list subjected to word segmentation into a conditional random field part-of-speech tagging model, and outputting a word list tagged with part-of-speech;
Inputting the word list marked with the part of speech into a dependency syntax analysis model, and outputting a word list containing the dependency relationship of each word;
inputting the word list containing the word dependency relationship into a distribution network scheduling semantic role identification model, and outputting a word list marked with distribution network scheduling semantic role attributes;
Inputting the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic identification model, and outputting key information of operation tasks in different types of maintenance application forms;
converting the equipment starting state and the termination state in the key information into an environment initial state matrix and an environment termination state matrix;
Inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to perform step-by-step reasoning, so as to obtain all ordered operation items corresponding to the operation task;
Inputting the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model, and outputting key information of operation tasks in different types of maintenance application forms, wherein the method specifically comprises the following steps of:
In a word list marked with distribution network scheduling semantic role attributes, performing paraphrasing search on the content of the operation type to determine the type of the overhaul application form;
Extracting station names, feeder line names, equipment names related to operation and equipment state names from a word list marked with distribution network scheduling semantic role attributes according to the type of the overhaul application form; searching a mediator relation in the dependency relation, and searching for a device name of a starting position and a terminating position of a power outage section or an operation section, a device name of a new device start/exit and a state name before and after the device operation from the device names related to the operation; searching parallel relations in the dependency relation, dividing sentences containing the parallel relations into a plurality of short sentences, and obtaining key information of an overhaul application form operation task;
The step-by-step reasoning is carried out by inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to obtain all ordered operation items corresponding to the operation task, and the method specifically comprises the following steps:
Inputting the environment initial state matrix into a deep learning reasoning network model, outputting a first operation item state by the model, and comparing the input environment initial state matrix with the output change of the first operation item state to obtain a first operation item; continuously inputting the first operation item state into a deep learning reasoning network model, outputting to obtain a second operation item state, and comparing the input first operation item state with the output second operation item state change to obtain a second operation item; and analogically, until the operation item state output by the model is the same as the environment termination state matrix, and obtaining all ordered operation items corresponding to the operation task;
The deep learning reasoning network model is trained by adopting a deep deterministic strategy gradient algorithm, and the training method of the deep learning reasoning network model comprises the following steps:
Acquiring a plurality of historical operation tickets;
For each history operation ticket, calling a ring network topological structure according to the contained equipment ID;
Based on a ring network topological structure, converting the initial state and the termination state of the equipment corresponding to each historical operation ticket into an environment initial state matrix X 0 and an environment termination state matrix X e;
And inputting an environment initial state matrix X 0 corresponding to each historical operation ticket and states X 1、X2、…、XT of all operation items, wherein T is the number of the operation items, respectively inputting a deep learning reasoning network model, sequentially outputting a variable 1 and a variable 2 … variable T by the model, and training the deep learning reasoning network model by taking the states that the variable 1 is the same as the state X 1 of the first operation item, the variable 2 is the same as the state X 2 of the second operation item, … and the variable T is the same as the state matrix X e of the environment termination.
2. The method for rapidly generating operation tickets for a power distribution network according to claim 1, wherein the conditional random field word segmentation model is trained by using a quasi-newton algorithm, and the training corpus comprises: beijing university word segmentation corpus, ICWB conference word segmentation corpus, MSR05 word segmentation corpus and related corpus for power distribution network scheduling of an electric power system.
3. The rapid generation method of the operation ticket for the power distribution network according to claim 1, wherein the conditional random field part-of-speech tagging model is trained by a viterbi algorithm, the part-of-speech tagging standard is "modern Chinese corpus processing Specification—word segmentation and part-of-speech tagging standard" of Beijing university, and the part-of-speech is divided by a coarse granularity standard.
4. The rapid generation method of the power distribution network operation ticket according to claim 1, wherein the dependency syntax analysis model is trained by a neural network algorithm, the corpus used for training comprises a semantic dependency network corpus tree library of Qinghai university and related corpora of power system power distribution network scheduling, each sentence is marked according to CONLL format, and marking content comprises sequence numbers, words, prototypes, coarse-granularity parts of speech, fine-granularity parts of speech, syntax features, center words and dependency relations.
5. The rapid generation method of a power distribution network operation ticket according to claim 1, wherein the power distribution network scheduling semantic role identification model comprises a bidirectional long-short-time memory network and a conditional random field layer connected with the bidirectional long-short-time memory network, and the marking method of semantic role attributes is as follows: and carrying out one-hot coding according to the positions of the words in the distribution network scheduling initial dictionary, and taking the coding as the label of the distribution network scheduling semantic roles.
6. The method for rapidly generating a power distribution network operation ticket according to claim 1, wherein the converting the device start state and the termination state in the key information into an environment initial state matrix and an environment termination state matrix specifically comprises:
And calling a ring network topological structure according to the equipment ID in the key information, and combining the equipment starting state and the termination state in the key information with the ring network topological structure to obtain an environment initial state matrix and an environment termination state matrix.
7. The utility model provides a distribution network ticket rapid generation device which characterized in that includes:
The word segmentation module is configured to input the work content, the type of the overhaul ticket and the sentence text in the stop or power transmission range rail in the overhaul application form into the conditional random field word segmentation model and output a word list after word segmentation;
the part-of-speech tagging module is configured to input the word list subjected to word segmentation into a conditional random field part-of-speech tagging model and output a word list tagged with part-of-speech;
The dependency syntax analysis module is configured to input the word list marked with the part of speech into a dependency syntax analysis model and output a word list containing the dependency relation of each word;
The semantic role recognition module is configured to input the word list containing the word dependency relationship into a distribution network scheduling semantic role recognition model and output a word list marked with distribution network scheduling semantic role attributes;
The semantic recognition module is configured to input the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model and output key information of operation tasks in different types of maintenance application forms;
the preprocessing module is configured to convert the equipment starting state and the equipment ending state in the key information into an environment initial state matrix and an environment ending state matrix;
The reasoning module is configured to input the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to conduct step-by-step reasoning so as to obtain all ordered operation items corresponding to the operation task;
Inputting the word list marked with the distribution network scheduling semantic role attribute into a maintenance application form semantic recognition model, and outputting key information of operation tasks in different types of maintenance application forms, wherein the method specifically comprises the following steps of:
In a word list marked with distribution network scheduling semantic role attributes, performing paraphrasing search on the content of the operation type to determine the type of the overhaul application form;
Extracting station names, feeder line names, equipment names related to operation and equipment state names from a word list marked with distribution network scheduling semantic role attributes according to the type of the overhaul application form; searching a mediator relation in the dependency relation, and searching for a device name of a starting position and a terminating position of a power outage section or an operation section, a device name of a new device start/exit and a state name before and after the device operation from the device names related to the operation; searching parallel relations in the dependency relation, dividing sentences containing the parallel relations into a plurality of short sentences, and obtaining key information of an overhaul application form operation task;
The step-by-step reasoning is carried out by inputting the environment initial state matrix and the environment termination state matrix into a deep learning reasoning network model to obtain all ordered operation items corresponding to the operation task, and the method specifically comprises the following steps:
Inputting the environment initial state matrix into a deep learning reasoning network model, outputting a first operation item state by the model, and comparing the input environment initial state matrix with the output change of the first operation item state to obtain a first operation item; continuously inputting the first operation item state into a deep learning reasoning network model, outputting to obtain a second operation item state, and comparing the input first operation item state with the output second operation item state change to obtain a second operation item; and analogically, until the operation item state output by the model is the same as the environment termination state matrix, and obtaining all ordered operation items corresponding to the operation task;
The deep learning reasoning network model is trained by adopting a deep deterministic strategy gradient algorithm, and the training method of the deep learning reasoning network model comprises the following steps:
Acquiring a plurality of historical operation tickets;
For each history operation ticket, calling a ring network topological structure according to the contained equipment ID;
Based on a ring network topological structure, converting the initial state and the termination state of the equipment corresponding to each historical operation ticket into an environment initial state matrix X 0 and an environment termination state matrix X e;
And inputting an environment initial state matrix X 0 corresponding to each historical operation ticket and states X 1、X2、…、XT of all operation items, wherein T is the number of the operation items, respectively inputting a deep learning reasoning network model, sequentially outputting a variable 1 and a variable 2 … variable T by the model, and training the deep learning reasoning network model by taking the states that the variable 1 is the same as the state X 1 of the first operation item, the variable 2 is the same as the state X 2 of the second operation item, … and the variable T is the same as the state matrix X e of the environment termination.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110994717.2A CN113792538B (en) | 2021-08-27 | 2021-08-27 | Method and device for rapidly generating operation ticket of power distribution network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110994717.2A CN113792538B (en) | 2021-08-27 | 2021-08-27 | Method and device for rapidly generating operation ticket of power distribution network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113792538A CN113792538A (en) | 2021-12-14 |
| CN113792538B true CN113792538B (en) | 2024-05-14 |
Family
ID=79182284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110994717.2A Active CN113792538B (en) | 2021-08-27 | 2021-08-27 | Method and device for rapidly generating operation ticket of power distribution network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113792538B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117093675A (en) * | 2023-10-18 | 2023-11-21 | 江苏泰坦智慧科技有限公司 | Operation ticket searching and matching method, device and storage medium |
| CN117875948B (en) * | 2024-03-12 | 2024-05-14 | 江苏秉信科技有限公司 | Equipment-based power grid overhaul event determination method, device, equipment and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487182A (en) * | 2019-09-12 | 2021-03-12 | 华为技术有限公司 | Training method of text processing model, and text processing method and device |
| KR20210058701A (en) * | 2019-11-13 | 2021-05-24 | 서강대학교산학협력단 | System and method for dependent parsing |
| CN112860872A (en) * | 2021-03-17 | 2021-05-28 | 广东电网有限责任公司 | Self-learning-based method and system for verifying semantic compliance of power distribution network operation tickets |
| CN113011771A (en) * | 2021-03-31 | 2021-06-22 | 广东电网有限责任公司 | Rapid ticketing method, device, equipment and storage medium |
| CN113159443A (en) * | 2021-04-30 | 2021-07-23 | 贵州电网有限责任公司 | Optimal path selection method and system suitable for operation order |
-
2021
- 2021-08-27 CN CN202110994717.2A patent/CN113792538B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487182A (en) * | 2019-09-12 | 2021-03-12 | 华为技术有限公司 | Training method of text processing model, and text processing method and device |
| KR20210058701A (en) * | 2019-11-13 | 2021-05-24 | 서강대학교산학협력단 | System and method for dependent parsing |
| CN112860872A (en) * | 2021-03-17 | 2021-05-28 | 广东电网有限责任公司 | Self-learning-based method and system for verifying semantic compliance of power distribution network operation tickets |
| CN113011771A (en) * | 2021-03-31 | 2021-06-22 | 广东电网有限责任公司 | Rapid ticketing method, device, equipment and storage medium |
| CN113159443A (en) * | 2021-04-30 | 2021-07-23 | 贵州电网有限责任公司 | Optimal path selection method and system suitable for operation order |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113792538A (en) | 2021-12-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113792538B (en) | Method and device for rapidly generating operation ticket of power distribution network | |
| CN110175324B (en) | Power grid operation instruction verification method and system based on data mining | |
| CN110287489A (en) | Document creation method, device, storage medium and electronic equipment | |
| CN112527997B (en) | Intelligent question-answering method and system based on power grid field scheduling scene knowledge graph | |
| CN110188345B (en) | Intelligent identification method and device for electric operation ticket | |
| CN112100397A (en) | Electric power plan knowledge graph construction method and system based on bidirectional gating circulation unit | |
| CN113095050A (en) | Intelligent ticketing method, system, equipment and storage medium | |
| CN112632288A (en) | Power dispatching system and method based on knowledge graph | |
| CN115357726A (en) | Fault disposal plan digital model establishing method based on knowledge graph | |
| CN114282010A (en) | Power grid operation fault identification method and system based on knowledge graph and storage medium | |
| CN112036179B (en) | Electric power plan information extraction method based on text classification and semantic frame | |
| CN113312924A (en) | Risk rule classification method and device based on NLP high-precision analysis label | |
| CN115544773A (en) | Method and system for reconstructing inaccurate measurement model, computer equipment and storage medium | |
| CN112417852A (en) | Method and device for judging importance of code segment | |
| CN110782145A (en) | Power dispatching plan decision making system, method and equipment | |
| CN116521889A (en) | Deep learning-based power grid dispatching comprehensive decision determining method and system | |
| CN115481964A (en) | Intelligent ticket forming method and device | |
| CN109800423A (en) | Method and apparatus are determined based on the power-off event of power failure plan sentence | |
| CN115099232A (en) | Power grid operation knowledge model construction method for scheduling regulations and historical data | |
| CN114996494A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN113010682A (en) | Command ticket system checking method, device and storage medium | |
| CN112860872B (en) | Power distribution network operation ticket semantic compliance verification method and system based on self-learning | |
| Sun et al. | Ensemble Machine Learning Identification of Power Fault Countermeasure Text Considering Word String TF-IDF Feature | |
| CN114997168B (en) | Power grid fault handling plan analysis method based on neuronization regular expression | |
| CN114186688A (en) | Maintenance application structured analysis method based on topology and service characteristic points |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |