CN111737993B - Method for extracting equipment health state from fault defect text of power distribution network equipment - Google Patents
Method for extracting equipment health state from fault defect text of power distribution network equipment Download PDFInfo
- Publication number
- CN111737993B CN111737993B CN202010455039.8A CN202010455039A CN111737993B CN 111737993 B CN111737993 B CN 111737993B CN 202010455039 A CN202010455039 A CN 202010455039A CN 111737993 B CN111737993 B CN 111737993B
- Authority
- CN
- China
- Prior art keywords
- fault defect
- text
- fault
- texts
- training set
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 131
- 230000036541 health Effects 0.000 title claims abstract description 54
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 46
- 238000012360 testing method Methods 0.000 claims abstract description 43
- 230000011218 segmentation Effects 0.000 claims abstract description 17
- 230000003862 health status Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 abstract description 4
- 238000013507 mapping Methods 0.000 abstract description 3
- 238000012423 maintenance Methods 0.000 description 4
- 238000005065 mining Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000000007 visual effect Effects 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
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Databases & Information Systems (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Probability & Statistics with Applications (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a method for extracting the health state of equipment from fault defect text of power distribution network equipment. Performing word segmentation and vectorization on fault defect text of power distribution network equipment; constructing a training set and a testing set; calculating the similarity between two fault defect texts in the test set and the training set; calculating the optimal value number k of the k-nearest neighbor algorithm, testing each fault defect text in a set, and selecting k fault defect texts with the maximum similarity in a training set; and calculating the health states of the fault defect texts in the test set, taking the health states of the fault defect texts in the test set as weighted average sums of the health states of k fault defect texts, and then carrying out weighted summation to obtain the final health state. According to the invention, self-learning mapping from fault defect text to health state data is realized, the existing fault/defect grade evaluation mode is changed, the health state of the equipment is extracted from the fault defect text based on a k-nearest neighbor algorithm, and the data of the health state of the whole equipment is accurately obtained.
Description
Technical Field
The invention belongs to a method for combining natural language processing and power distribution network equipment data information in the field of intelligent power operation and detection, and particularly relates to a method for extracting equipment health state from fault defect text of power distribution network equipment.
Background
With the acceleration of the grid informatization process, many unstructured and semi-structured data are accumulated in the database of the grid enterprise. As one of the most typical, most complex unstructured data, analysis of text-type data has been a hotspot problem in the field of data mining.
In the overhaul and maintenance link of the power system, a large number of overhaul test records, inspection defect elimination records, fault and defect description reports, event sequence records and the like are recorded. The logs and reports mainly show text for short in Chinese text mixed with numbers and alphabetic characters), and are rich in equipment history running state information, overhaul effect information, reliability information and the like, and have great buddhist benefits in the objective evaluation of the development process of the health state of the equipment.
However, since texts have characteristics of multiple ambiguity, difficulty in segmentation, ambiguity, multiple noise, etc., the above information has not been fully mined yet. The Chinese text processing of the power grid belongs to a starting stage at present. Visual reliability statistical information is mined from Chinese text information, and complex information mining technology and smart mining processes need to be explored. Chinese text mining has long been considered an important and difficult technique. Especially when it is applied to various professional fields, it is more difficult to closely combine knowledge in the professional fields. In the power field, foreign students propose to excavate massive historical defect data by using a machine learning method aiming at a New York power grid, so that the basis of power equipment fault prediction and preventive maintenance is provided. However, chinese text is very different from english text, and not only there are no spaces between words, but also the part of speech and the syntax structure are very different, and the processing mode of carrying english text is not feasible. Other texts in the power system, such as fault text, are manually typed, have complex syntactic structures, and are difficult to accurately divide main guest components, so that the processing is complex and difficult.
The equipment defect classification and analysis statistics work to be performed by the power grid enterprises every year is often performed manually, so that the workload is large, time and labor are consumed, and the correctness of the classification and statistics work is difficult to verify due to subjective factors and experience differences. Data analysis mining based on fault defect text is therefore important and urgent.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and discloses a mapping method based on fault defect text to equipment health state evaluation.
The technical scheme adopted for solving the technical problems is as follows:
the method comprises the following steps:
1) Word segmentation is carried out on each fault defect text of the power distribution network equipment, and word segmentation results are obtained;
the fault defect text refers to text data which are input into a computer by different maintenance/repair staff at different times for power distribution network equipment of a power grid system.
2) Vectorizing fault defect text;
3) Constructing a training set and a testing set of fault defect texts of power distribution network equipment;
collecting all fault defect texts of known power distribution network equipment, carrying out vectorization processing according to the method of the step 2), classifying according to health states, forming a fault defect text library, and taking the fault defect text library as a training set; vectorizing all fault defect texts of the power distribution network equipment to be tested according to the method of the step 2) to form a test set;
4) Calculating the similarity between every two fault defect texts in the test set and the training set;
5) Calculating the optimal value of the optimal value number k in the k-nearest neighbor algorithm, and then selecting k fault defect texts with the highest similarity from the training set aiming at each fault defect text in the test set;
in the conventional method, the value of the parameter k is preset. According to the training set of the fault defect text, the optimal k value is calculated, so that the prediction accuracy rate on the training set is highest.
6) Calculating a health state result HI (Health Index) of each fault defect text in the test set, wherein HI is a decimal between 0 and 1, namely HI is more than or equal to 0 and less than or equal to 1,0 and 1 respectively represent equipment faults and complete health, and taking the health state of each fault defect text in the test set as a weighted average sum of health states corresponding to k fault defect texts in the training set obtained in the step (5); and then, carrying out weighted summation on the health states of the fault defect texts in the test set to obtain the final health state H of the power distribution network equipment.
The step 5) is specifically as follows:
5.1 Dividing the health state HI into 3 types, wherein the corresponding area areas are [0,0.33 ], [0.33,0.67 ], [0.67,1], the initial value of the optimal value number k is 20, and the initial value of the selected fault defect text number n of the fault defect texts in the training set is set;
5.2 Aiming at the vectorized results of n fault defect texts in the training set, n=0 to J, and selecting k fault defect texts with highest similarity from the training set according to the similarity obtained by calculation in the step 4);
5.3 Taking the categories of the health states HI corresponding to the k fault defect texts, carrying out average calculation to obtain prediction categories, and then judging:
if the corresponding category of the fault defect text in the training set is consistent with the predicted category, the fault defect text is considered to be correct, otherwise, the fault defect text is considered to be wrong;
5.4 Counting the correct rate corresponding to n fault defect texts;
5.5 If the correct rate of the current iteration is obviously increased by 20% or more than that of the previous iteration, the k value is reduced by 1, and the step returns to the step 5.2), and the step process is repeated;
if the correct rate of the current iteration is not obviously increased by 20% or more than that of the previous iteration, the current k value is an optimal value;
5.6 Aiming at each fault defect text in the test set, selecting k fault defect texts with the largest similarity in the training set.
In the step 6), the weighted average sum of the health states corresponding to the k fault defect texts in the training set obtained in the step (5) is used as the health state of each fault defect text in the test set by adopting the following formula:
wherein HI x Representing the health state of an xth fault defect sample in k fault defect texts selected in the training set;indicating the selected w of the test set i Health status of individual fault defect samples.
The step 1) adopts a Chinese word segmentation technology (HMM-CWS) method based on a hidden Markov model to perform preprocessing of text word segmentation.
Step 2) is to sort the one-dimension of each word corresponding to the vector space after word segmentation according to word frequency, remove repeated words, obtain non-repeated word sequences, and the non-repeated word sequences form a complete vector space.
And 4) calculating the similarity between each fault defect text in the test set and each fault defect text in the training set according to a similarity measurement formula.
The invention has the beneficial effects that:
according to the invention, self-learning mapping from fault defect text to health state data is realized, the existing fault/defect grade evaluation mode is changed, the health state of the equipment is extracted from the fault defect text based on a k-nearest neighbor algorithm, and a basis is provided for the power grid enterprise to accurately obtain the data of the health state of the whole equipment.
Drawings
Fig. 1 is a flow chart of the method of the present invention.
Detailed Description
The invention is further described below with reference to the drawings and examples.
The embodiment of the invention and the implementation process are as follows:
the fault condition of the power distribution network equipment of the power grid system is recorded at different times by different maintenance/repair staff, and the fault condition is input into a computer in the form of text to form fault defect text data of the power distribution network equipment.
1) Word segmentation is carried out on each fault defect text of the power distribution network equipment, and word segmentation results are obtained;
specifically, a Chinese word segmentation technology (HMM-CWS) method based on a hidden Markov model is adopted for preprocessing text word segmentation.
2) Vectorizing fault defect text;
step 1) is to sort the one-dimension of each word corresponding to the vector space after word segmentation according to word frequency, remove repeated words, obtain non-repeated word sequences, and the non-repeated word sequences form a complete vector space.
In particular, the formation of the complete vector space W is shown as follows ALL :
W ALL =[w ij ] IxJ
Wherein w is ij Representing the weight, w, between the ith fault defect text in the test set and the jth fault defect text in the training set ij =0 or 1. When w is ij When=1, the text contains the word vector, otherwise 0; i represents the total number of fault defect texts in the test set, J represents the total number of fault defect texts in the training set, [] IxJ The size of the representation matrix is i×j.
3) Constructing a training set and a testing set of fault defect texts of power distribution network equipment;
collecting all fault defect texts of known power distribution network equipment, carrying out vectorization processing according to the method of the step 2), classifying according to the health state to form a fault defect text library, and taking the fault defect text library as a training set, wherein each fault defect text corresponds to a specific health state classification; vectorizing all fault defect texts of the power distribution network equipment to be tested according to the method of the step 2) to form a test set; the test set is a vectorized set of fault defect text to be classified.
In particular, the classification according to health status is specifically classified into three categories of complete health, sub-health and failure.
4) Calculating the similarity between every two fault defect texts in the test set and the training set;
specifically, the similarity between each fault defect text in the test set and each fault defect text in the training set is calculated according to a similarity measurement formula.
The similarity formula is as follows:
wherein S is ij For the similarity between the ith fault defect text in the test set and the jth fault defect text in the training set, w i Is the feature vector of text i, w j Is the feature vector of text j, M is the dimension of the vector, w il 、w jl Is the vector w i 、w j Is a first dimension value of (c).
5) Calculating the optimal value of the optimal value number k in the k-nearest neighbor algorithm, and then selecting k fault defect texts with the highest similarity from the training set aiming at each fault defect text in the test set;
5.1 Health Index, HI is a decimal between 0 and 1, i.e., 0.ltoreq.HI.ltoreq.1, 0 and 1 represent equipment failure and complete Health, respectively) are classified into 3 classes, corresponding zones are [0,0.33 ], [0.33,0.67 ], [0.67,1], and the initial value of the optimal number k is 20;
5.2 Aiming at the vectorized result of n fault defect texts in the training set, the value range of n is an integer of [1, J ], and k fault defect texts with highest similarity are selected from the training set according to the similarity obtained by calculation in the step 4);
5.3 Taking the categories of the health states HI corresponding to the k fault defect texts, carrying out average calculation to obtain prediction categories, and then judging:
if the corresponding category of the fault defect text in the training set is consistent with the predicted category, the fault defect text is considered to be correct, otherwise, the fault defect text is considered to be wrong;
5.4 Counting the correct rate corresponding to n fault defect texts;
5.5 If the accuracy of the current iteration is increased by 20% or more than that of the previous iteration, the k value is reduced by 1, and the step is returned to the step 5.2), and the step process is repeated;
if the correct rate of the current iteration is not increased by 20% or more than that of the previous iteration, the current k value is an optimal value;
5.6 Aiming at each fault defect text in the test set, selecting k fault defect texts with the largest similarity in the training set.
6) Calculating health state results HI of each fault defect text in the test set, and taking the health state of each fault defect text in the test set as a weighted average sum of health states corresponding to k fault defect texts in the training set obtained in the step (5); and then, carrying out weighted summation on the health states of the fault defect texts in the test set to obtain the final health state HI of the power distribution network equipment.
Specifically, the weighted average sum of health states corresponding to k fault defect texts in the training set obtained in the step (5) is used as the health state of each fault defect text in the test set by adopting the following formula:
wherein HI x Representing the health state of an xth fault defect sample in k fault defect texts selected in the training set;indicating the selected w of the test set i Health status of individual fault defect samples.
Claims (2)
1. A method for extracting the health state of equipment by fault defect text of power distribution network equipment is characterized by comprising the following steps: the method comprises the following steps:
1) Word segmentation is carried out on each fault defect text of the power distribution network equipment, and word segmentation results are obtained;
2) Vectorizing fault defect text;
3) Constructing a training set and a testing set of fault defect texts of power distribution network equipment;
collecting all fault defect texts of known power distribution network equipment, carrying out vectorization processing according to the method of the step 2), classifying according to health states, forming a fault defect text library, and taking the fault defect text library as a training set; vectorizing all fault defect texts of the power distribution network equipment to be tested according to the method of the step 2) to form a test set;
4) Calculating the similarity between every two fault defect texts in the test set and the training set;
5) Calculating the optimal value of the optimal value number k in the k-nearest neighbor algorithm, and then selecting k fault defect texts with the highest similarity from the training set aiming at each fault defect text in the test set;
6) Calculating the health status result HI of each fault defect text in the test set, wherein HI is a decimal from 0 to 1, namely HI is more than or equal to 0 and less than or equal to 1,0 and 1 respectively represent equipment faults and complete health, and taking the health status of each fault defect text in the test set as the weighted average sum of the health statuses corresponding to k fault defect texts in the training set obtained in the step (5); then, carrying out weighted summation on the health states of the fault defect texts in the test set to obtain a final health state H of the power distribution network equipment;
the step 5) is specifically as follows:
5.1 Dividing the health state HI into 3 types, wherein the corresponding area areas are [0,0.33 ], [0.33,0.67 ], [0.67,1], the initial value of the optimal value number k is 20, and the initial value of the selected fault defect text number n of the fault defect texts in the training set is set;
5.2 Aiming at the vectorized results of n fault defect texts in the training set, n=0 to J, and selecting k fault defect texts with highest similarity from the training set according to the similarity obtained by calculation in the step 4);
5.3 Taking the categories of the health states HI corresponding to the k fault defect texts, carrying out average calculation to obtain prediction categories, and then judging:
if the corresponding category of the fault defect text in the training set is consistent with the predicted category, the fault defect text is considered to be correct, otherwise, the fault defect text is considered to be wrong;
5.4 Counting the correct rate corresponding to n fault defect texts;
5.5 If the correct rate of the current iteration is obviously increased by 20% or more than that of the previous iteration, the k value is reduced by 1, and the step returns to the step 5.2), and the step process is repeated;
if the correct rate of the current iteration is not obviously increased by 20% or more than that of the previous iteration, the current k value is an optimal value;
5.6 Aiming at each fault defect text in the test set, selecting k fault defect texts with the maximum similarity in the training set;
in the step 6), the weighted average sum of the health states corresponding to the k fault defect texts in the training set obtained in the step (5) is used as the health state of each fault defect text in the test set by adopting the following formula:
wherein HI x Representing the health state of an xth fault defect sample in k fault defect texts selected in the training set;indicating the selected w of the test set i Health status of individual fault defect samples;
the method comprises the following steps of 1) preprocessing text word segmentation by adopting a Chinese word segmentation technical method based on a hidden Markov model;
step 2) is to sort the one-dimension of each word corresponding to the vector space after word segmentation according to word frequency, remove repeated words, obtain non-repeated word sequences, and the non-repeated word sequences form a complete vector space.
2. The method for extracting the health state of the power distribution network equipment by using the fault defect text according to claim 1, wherein the method comprises the following steps of: and 4) calculating the similarity between each fault defect text in the test set and each fault defect text in the training set according to a similarity measurement formula.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010455039.8A CN111737993B (en) | 2020-05-26 | 2020-05-26 | Method for extracting equipment health state from fault defect text of power distribution network equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010455039.8A CN111737993B (en) | 2020-05-26 | 2020-05-26 | Method for extracting equipment health state from fault defect text of power distribution network equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111737993A CN111737993A (en) | 2020-10-02 |
CN111737993B true CN111737993B (en) | 2024-04-02 |
Family
ID=72647680
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010455039.8A Active CN111737993B (en) | 2020-05-26 | 2020-05-26 | Method for extracting equipment health state from fault defect text of power distribution network equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111737993B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112580875B (en) * | 2020-12-21 | 2022-06-24 | 泉州装备制造研究所 | Fault prediction method and system for power distribution device |
CN112925668B (en) * | 2021-02-25 | 2024-04-05 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for evaluating server health |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105303296A (en) * | 2015-09-29 | 2016-02-03 | 国网浙江省电力公司电力科学研究院 | Electric power equipment full-life state evaluation method |
CN105514991A (en) * | 2015-12-08 | 2016-04-20 | 华北电力科学研究院有限责任公司 | New energy power station generating equipment health state determining method and apparatus |
WO2017167067A1 (en) * | 2016-03-30 | 2017-10-05 | 阿里巴巴集团控股有限公司 | Method and device for webpage text classification, method and device for webpage text recognition |
CN109977228A (en) * | 2019-03-21 | 2019-07-05 | 浙江大学 | The information identification method of grid equipment defect text |
WO2019200806A1 (en) * | 2018-04-20 | 2019-10-24 | 平安科技(深圳)有限公司 | Device for generating text classification model, method, and computer readable storage medium |
CN110895565A (en) * | 2019-11-29 | 2020-03-20 | 国网湖南省电力有限公司 | Method and system for classifying fault defect texts of power equipment |
-
2020
- 2020-05-26 CN CN202010455039.8A patent/CN111737993B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105303296A (en) * | 2015-09-29 | 2016-02-03 | 国网浙江省电力公司电力科学研究院 | Electric power equipment full-life state evaluation method |
CN105514991A (en) * | 2015-12-08 | 2016-04-20 | 华北电力科学研究院有限责任公司 | New energy power station generating equipment health state determining method and apparatus |
WO2017167067A1 (en) * | 2016-03-30 | 2017-10-05 | 阿里巴巴集团控股有限公司 | Method and device for webpage text classification, method and device for webpage text recognition |
WO2019200806A1 (en) * | 2018-04-20 | 2019-10-24 | 平安科技(深圳)有限公司 | Device for generating text classification model, method, and computer readable storage medium |
CN109977228A (en) * | 2019-03-21 | 2019-07-05 | 浙江大学 | The information identification method of grid equipment defect text |
CN110895565A (en) * | 2019-11-29 | 2020-03-20 | 国网湖南省电力有限公司 | Method and system for classifying fault defect texts of power equipment |
Non-Patent Citations (2)
Title |
---|
文本信息挖掘技术及其在断路器全寿命状态评价中的应用;邱剑;王慧芳;应高亮;张波;邹国平;何奔腾;;电力系统自动化(06);正文第107-118页 * |
考虑历史缺陷文本信息的断路器状态评价研究;马润泽;王龙响;余佳文;王慧芳;邱剑;;机电工程(10);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111737993A (en) | 2020-10-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111882446B (en) | Abnormal account detection method based on graph convolution network | |
CN108985380B (en) | Point switch fault identification method based on cluster integration | |
CN111709244B (en) | Deep learning method for identifying cause and effect relationship of contradictory dispute | |
CN110175324B (en) | Power grid operation instruction verification method and system based on data mining | |
CN111427775B (en) | Method level defect positioning method based on Bert model | |
CN108304567B (en) | Method and system for identifying working condition mode and classifying data of high-voltage transformer | |
CN111274814B (en) | Novel semi-supervised text entity information extraction method | |
CN111737993B (en) | Method for extracting equipment health state from fault defect text of power distribution network equipment | |
CN111340110B (en) | Fault early warning method based on industrial process running state trend analysis | |
CN101738998B (en) | System and method for monitoring industrial process based on local discriminatory analysis | |
CN109101483A (en) | A kind of wrong identification method for electric inspection process text | |
CN112818008A (en) | Intelligent diagnosis method, system, medium and electronic equipment for nuclear power debugging faults | |
CN113342784A (en) | Database design method for risk assessment of main transformer equipment of power grid | |
CN104281525A (en) | Defect data analytical method and method for shortening software testing programs by using same | |
CN113435759A (en) | Primary equipment risk intelligent evaluation method based on deep learning | |
CN113065356A (en) | IT equipment operation and maintenance fault suggestion processing method based on semantic analysis algorithm | |
CN114091549A (en) | Equipment fault diagnosis method based on deep residual error network | |
CN114580978A (en) | System and method for inspecting quality of ring-comment report | |
CN116432123A (en) | Electric energy meter fault early warning method based on CART decision tree algorithm | |
CN114662405A (en) | Rock burst prediction method based on few-sample measurement and ensemble learning | |
CN107577738A (en) | A kind of FMECA method by SVM text mining processing datas | |
CN114491081A (en) | Electric power data tracing method and system based on data blood relationship graph | |
CN112559741A (en) | Nuclear power equipment defect recording text classification method, system, medium and electronic equipment | |
CN111209955A (en) | Airplane power supply system fault identification method based on deep neural network and random forest | |
CN114880584B (en) | Generator set fault analysis method based on community discovery |
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 |