WO2024103436A1 - 设备缺陷数据规则库构建方法及设备缺陷关联性分析方法 - Google Patents

设备缺陷数据规则库构建方法及设备缺陷关联性分析方法 Download PDF

Info

Publication number
WO2024103436A1
WO2024103436A1 PCT/CN2022/134353 CN2022134353W WO2024103436A1 WO 2024103436 A1 WO2024103436 A1 WO 2024103436A1 CN 2022134353 W CN2022134353 W CN 2022134353W WO 2024103436 A1 WO2024103436 A1 WO 2024103436A1
Authority
WO
WIPO (PCT)
Prior art keywords
defect
equipment
data
feature
equipment defect
Prior art date
Application number
PCT/CN2022/134353
Other languages
English (en)
French (fr)
Inventor
肖耀辉
黄和燕
周震震
余俊松
王玉峰
李为明
罗征洋
何钰
宋云海
何森
黄怀霖
梁皓云
Original Assignee
中国南方电网有限责任公司超高压输电公司检修试验中心
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国南方电网有限责任公司超高压输电公司检修试验中心 filed Critical 中国南方电网有限责任公司超高压输电公司检修试验中心
Publication of WO2024103436A1 publication Critical patent/WO2024103436A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • 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
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/2431—Multiple classes
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Definitions

  • the present application relates to the field of power transmission safety technology, and in particular to a method, device, computer device and storage medium for constructing a device defect data rule base, and a method, device, computer device and storage medium for analyzing the correlation of device defects.
  • the normal opening, closing and status monitoring of the converter valve control equipment plays a vital role in the stable operation of the DC system.
  • the normal opening, closing and status monitoring of the converter valve control equipment plays a vital role in the stable operation of the DC system.
  • valve-based electronic equipment often fails, and the types of failures and causes of failures are also different.
  • some faults can only be handled after the DC system is shut down, and most fault defects are manually eliminated and recorded by operation and maintenance personnel, without in-depth analysis of the causes of the faults.
  • a method, apparatus, computer equipment and storage medium for constructing a device defect data rule base, and a method, apparatus, computer equipment and storage medium for analyzing device defect correlation are provided.
  • a method for constructing a device defect data rule base comprising:
  • a method for analyzing the correlation of equipment defects comprising:
  • the equipment defect data rule base is constructed using the above-mentioned equipment defect data rule base construction method, and the equipment defect analysis results include defect influencing factors, faulty component modules and defect level prediction results.
  • a device for constructing a rule base for equipment defect data comprising:
  • a data acquisition module used to acquire original equipment defect data set
  • a data extraction module is used to extract device defect features from an original device defect data set to obtain a device defect feature set, wherein the device defect feature set includes defect features, defect influencing factors, and component modules;
  • the defect grading module is used to predict the defect grading of the equipment defect feature set and obtain the equipment defect grade prediction result;
  • a correlation analysis module is used to perform correlation analysis on the equipment defect feature set and the equipment defect level prediction result to obtain the correlation between the defect features, defect influencing factors and component modules;
  • the rule base construction module is used to construct the equipment defect data rule base based on the association relationship.
  • a device for analyzing the correlation of equipment defects comprising:
  • a data acquisition module used to acquire equipment defect data
  • a defect analysis module is used to match the equipment defect data in the established equipment defect data rule library to obtain equipment defect analysis results;
  • the equipment defect data rule base is constructed by using the equipment defect data rule base construction method as described above, and the equipment defect analysis results include defect influencing factors, faulty component modules, and defect level prediction results.
  • a computer device includes a memory and one or more processors, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the one or more processors execute the steps in the above-mentioned device defect data rule base construction method or the above-mentioned device defect correlation analysis method.
  • One or more computer storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to execute the steps in the above-mentioned device defect data rule base construction method or the above-mentioned device defect correlation analysis method.
  • FIG1 is an application environment diagram of a method for constructing a device defect data rule base or a method for analyzing device defect correlation in one embodiment
  • FIG2 is a schematic diagram of a process of constructing a device defect data rule base in one embodiment
  • FIG3 is a schematic flow chart of the step of extracting a device defect feature set in one embodiment
  • FIG4 is a schematic flow chart of the step of extracting a device defect feature set in another embodiment
  • FIG5 is a schematic diagram of a detailed flow chart of a method for constructing a device defect data rule base in one embodiment
  • FIG6 is a schematic diagram of a process flow of a method for analyzing the correlation of equipment defects in one embodiment
  • FIG7 is a schematic diagram of a flow chart of a step of matching device defect data in one embodiment
  • FIG8 is a structural block diagram of a device for constructing a rule base for equipment defect data in one embodiment
  • FIG9 is a structural block diagram of a device for analyzing equipment defect correlation in one embodiment
  • FIG. 10 is a diagram showing the internal structure of a computer device in one embodiment.
  • the method for constructing a device defect data rule base can be applied in an application environment as shown in FIG1 .
  • the terminal 102 communicates with the server 104 via a network.
  • the data storage system can store data that the server 104 needs to process.
  • the data storage system can be integrated on the server 104, or it can be placed on a cloud or other network server.
  • the operation and maintenance personnel can upload the original device defect data set to the server 104 through the terminal 102, and send a rule base construction message to the server 104.
  • the server 104 responds to the message, obtains the original device defect data set, extracts the device defect features of the original device defect data set, and obtains the device defect feature set.
  • the device defect feature set includes defect features, defect influencing factors, and component modules.
  • the defect classification prediction is performed on the device defect feature set to obtain the device defect level prediction result.
  • the device defect feature set and the device defect level prediction result are analyzed for correlation, and the correlation between the defect features, defect influencing factors, and component modules is obtained.
  • the device defect data rule base is constructed based on the correlation relationship.
  • the terminal 102 may be, but is not limited to, a personal computer, a laptop, a smart phone, a tablet computer, an IoT device, and a portable wearable device.
  • the IoT device may be a smart speaker, a smart TV, a smart air conditioner, a smart vehicle-mounted device, etc.
  • the portable wearable device may be a smart watch, a smart bracelet, a head-mounted device, etc.
  • the server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.
  • a method for constructing a device defect data rule base is provided, and the method is applied to the server in FIG1 as an example for description, including the following steps:
  • Step S100 obtaining an original equipment defect data set.
  • the original equipment defect data set refers to the defect data set of the equipment obtained through the preliminary data preparation work uploaded by the terminal.
  • the original equipment defect data set includes the defect type of the equipment, the initial defect level determined by humans, the influencing factors, the fault components, the fault description characteristics, the time information, and the weather conditions.
  • the original equipment defect data set can be obtained by the following methods: constructing a physical model of the valve-based electronic equipment, analyzing the composition and function of each component of the valve-based electronic equipment according to the principle of the DC transmission system, and analyzing the equipment defect characteristics and their influencing factors.
  • the defect data of the local valve-based electronic equipment is collected, and the relevant information such as the various defect alarms of the equipment and their influencing factors are summarized and analyzed according to the characteristics, and the multi-dimensional influencing factor system of the valve-based electronic equipment defects and the mapping relationship between the two are obtained to obtain the original equipment defect data set.
  • the physical model of the valve-based electronic equipment can be established, and the typical defects can be selected by classifying the defect influencing factors, including signal abnormality, optical power overload, excessive far-end bit errors, magnetic circuit abnormality, severe network packet loss, CPU (Central Processing Unit) frequency/voltage abnormality, reduced cooling efficiency, etc.
  • the local valve base electronic equipment defect data is collected. Each data is taken as a sample.
  • n data samples are selected to analyze the defect characteristics and their influencing factors, including fiber optic connector heating, pigtail interruption, instantaneous communication interference, software problems, equipment running time, internal communication interruption, improper human operation, external temperature changes and other influencing factors.
  • a multi-dimensional influencing factor system is generated to obtain the original equipment defect data set of valve base electronic equipment defects.
  • Step S200 extracting device defect features from an original device defect data set to obtain a device defect feature set, wherein the device defect feature set includes defect features, defect influencing factors, and component modules.
  • the amount of data in the original equipment defect data set is very large, and the data is disorganized. Therefore, after obtaining the original equipment defect feature set, the original equipment defect data set can be preprocessed by data cleaning, and then the features in the data set are evaluated based on the feature correlation function. Combined with the MapReduce framework, the features in the original equipment defect data set are extracted in parallel, including defect type features, defect influencing factors, and component modules.
  • the equipment defect feature set is obtained.
  • typical defect types of valve base electronic equipment include signal anomalies, optical power overload, excessive far-end bit errors, magnetic circuit anomalies, severe network packet loss, CPU frequency/voltage anomalies, and reduced cooling efficiency.
  • valve base electronic equipment By analyzing the performance and operating parameters of valve base electronic equipment, it is found that its defect influencing factors include heating of optical fiber connectors, pigtail interruption, instantaneous communication interference, software problems, equipment operation time, and internal communication interruption, and the component modules include optical transmitter boards, optical receiver boards, and signal relays.
  • Step S300 performing defect classification prediction on the equipment defect feature set to obtain an equipment defect grade prediction result.
  • the defect classification prediction can be performed on the equipment defect feature set to obtain the equipment defect grade prediction result. Specifically, the defect classification prediction can be performed on the equipment defect feature set through a classifier to obtain the equipment defect grade prediction result.
  • the defect classification can be divided into 5 levels, namely "very urgent", “urgent”, “major”, “general” and “other". It can be understood that in other embodiments, the defect grade can also be performed in other ways.
  • Step S400 performing correlation analysis on the equipment defect feature set and the equipment defect level prediction result to obtain the correlation relationship between the defect features, defect influencing factors and component modules.
  • the equipment defect feature set can be quantized and encoded, and then the equipment defect type, defect influencing factors and component modules are analyzed for correlation. Specifically, the features that affect the generation of defect alarms can be determined, and the correlation between defect influencing factors, defect type features and component modules is formed by calculating the support and confidence of the defect database. If the correlation between the three is very strong in the correlation, it is a strong correlation (hereinafter referred to as a strong rule), which indicates that when this type of defect occurs, the cause of the defect is largely related to the corresponding defect influencing factors in the correlation, and the faulty component is also likely to be the corresponding component module in the correlation.
  • a strong rule indicates that when this type of defect occurs, the cause of the defect is largely related to the corresponding defect influencing factors in the correlation, and the faulty component is also likely to be the corresponding component module in the correlation.
  • the number of correlations can be many, and the correlations can include one-to-many, many-to-one and many-to-many relationships.
  • Each correlation corresponds to a corresponding support and confidence to characterize the degree of correlation between defect type features, defect influencing factors and component modules.
  • defect influencing factors a, b, c and d can correspond to defect type feature A
  • component modules 1, 2, 3 and 4 can correspond to defect type feature A.
  • Step S500 constructing a device defect data rule base based on the association relationship.
  • association relationships can be integrated together to construct an equipment defect data rule base.
  • the association relationships can be integrated according to the confidence level to obtain an equipment defect rule database.
  • the equipment defect feature of the original equipment defect data set is extracted to obtain the equipment defect feature set, and the defect classification prediction is performed on the equipment defect feature set to obtain the equipment defect level prediction result, and then the equipment defect feature set and the equipment defect level prediction result are analyzed for correlation to obtain the defect features, defect influencing factors and the correlation between component modules, and finally, the equipment defect data rule base is constructed based on the correlation relationship.
  • the whole process, through the equipment defect data rule base can provide an accurate basis for the subsequent equipment defect correlation analysis, fault tracing and fault diagnosis, thereby improving the accuracy of fault diagnosis.
  • step 200 includes:
  • Step S202 constructing a feature evaluation function of the original equipment defect data set.
  • Step S204 evaluating the equipment defect features in the original equipment defect data set according to the feature evaluation function, and screening out an initial equipment defect feature set.
  • Step S206 construct a feature correlation function of the initial device defect feature set, and determine the correlation between any two features in the initial device defect feature set according to the feature correlation function.
  • Step S208 clustering the initial equipment defect feature set according to the correlation between each pair of features in the initial equipment defect feature set to obtain a clustering result.
  • Step S210 extracting equipment defect features from the clustering results to obtain an equipment defect feature set.
  • the feature judgment function may also be called a feature evaluation function, which is used to evaluate the correlation between features.
  • the feature correlation function is a function that evaluates the similarity between two features.
  • the feature judgment function F(x) may be constructed by: based on the equipment defect feature strategy of information theory, using the relevant knowledge of information theory to measure the degree of influence between defect features-labels and equipment defects-influencing factors, and constructing a feature judgment function. Then, the feature judgment function is used to evaluate the equipment defect features in the original equipment defect data set, and the feature corresponding to the maximum value of the feature judgment value F(i) is added to the initial equipment defect feature set F.
  • the above steps are iteratively performed until the number of features in F reaches a preset threshold, and the remaining minor features are automatically added to the feature set F'.
  • the feature set F' may not be processed or deleted. Then, the feature correlation function of the initial equipment defect feature set F is constructed:
  • xa and xb are any two different features in the initial equipment defect feature set F, and I( xa ; Y
  • the correlation between the two features in the initial equipment defect feature set is determined by the feature correlation function. The smaller the value of C( xa , xb ), the greater the similarity between the feature xa and the feature xb , and vice versa.
  • the initial equipment defect feature set can be clustered according to the correlation between the two features, and the similar features of the equipment defects are clustered in the same cluster.
  • I ⁇ I1 , I2 , ..., Ik ⁇ is the set of clusters returned after clustering.
  • the MapReduce framework is called to evenly distribute clusters according to the number of nodes in the framework.
  • Each node extracts features from the cluster using the PCA (Principle Component Analysis) algorithm by calling the Map function to obtain the device defect feature set G.
  • PCA Principal Component Analysis
  • the initial device defect feature set is screened out by the feature evaluation function, and the feature correlation function is used to extract features from the initial device defect feature set, so that the originally scattered and chaotic features become closely related, and connections are established between the features.
  • step S202 includes:
  • Step S222 determining the information gain of the equipment defect features in the original equipment defect data set, constructing an influencing factor function of the equipment defect features and influencing factors, and constructing a feature evaluation function of the original equipment defect data set based on the information gain and the influencing factor function.
  • Step S204 also includes: step S224, evaluating the device defect features in the original device defect data set according to the feature evaluation function, screening out the initial device defect feature set, screening out the target device defect features according to the information gain of the device defect features, and adding the target device defect features to the initial device defect feature set to update the initial device defect feature set.
  • the construction process of the feature evaluation function can be: preprocessing the original equipment defect data set, xi is any feature in the original equipment defect data set, including defect type features, influencing factor features, location features, component module features, etc.
  • the above features can be collectively referred to as equipment defect features, Y is the corresponding category label (i.e., the category of the defect), and the information gain G( xi ; Y) of the equipment defect feature xi is obtained as shown in formula (2).
  • H(Y) is the information entropy about the category label Y
  • xi ) is the conditional entropy about the equipment defect featurexi and category Y.
  • the influencing factor function Q between equipment defects and defect influencing factors is constructed, as shown in formula (3).
  • S is the original equipment defect dataset
  • Y is the category label
  • xj is the feature element in the dataset S.
  • I(x j ; Y) is the mutual information
  • xi ) is the conditional mutual information about the variable x j and the category label Y.
  • the feature judgment function F(i) of the original equipment defect data set is constructed, which can be specifically expressed as:
  • C is the weight parameter of the function G( xi ;Y).
  • the default file block strategy in Hadoop can be used to divide the feature space of the original data set into file blocks of the same size.
  • the file block Block is used as input data, and the Mapper node calls the Map function in the form of a key-value pair ⁇ key, value> to count the information gain of each feature (key is the feature name, value is the information gain of the corresponding feature), and combines each key-value pair to obtain a feature information gain set A.
  • the elements in set A are arranged in descending order according to the information gain value corresponding to the feature, and the features with a relatively low ranking in set A are removed, and a new defective device defect feature matrix X* is recombined.
  • the feature judgment function is constructed by information gain and influencing factor function, so that the feature judgment function can accurately measure the degree of influence between defect feature-label and equipment defect-influencing factor, thereby making the extracted device defect feature set more representative.
  • step S300 includes:
  • Step 302 perform word segmentation processing on the equipment defect description features in the equipment defect feature set to obtain a word segmentation result, determine the importance of the keywords in the word segmentation result, and based on the importance of the keywords, use a support vector machine to perform defect grade prediction on the equipment defect feature set to obtain an equipment defect grade prediction result.
  • the equipment defect feature set includes an equipment defect description document, which includes equipment defect description features.
  • the operation and maintenance personnel have pre-built a custom dictionary.
  • the defect classification of the equipment defect feature set can be: based on the custom dictionary, each equipment defect description feature is segmented, and then the TF-IDF algorithm is used to process each keyword after segmentation in the equipment defect description feature to determine the importance of the keyword in the segmentation result.
  • the calculation formula of TF-IDF is:
  • tf ij is the word frequency
  • ni,j is the number of times the description word ti appears in the defect description document dj
  • ⁇ k n k,j is the sum of all the description words in the defect description document dj.
  • idf i is the inverse document frequency
  • is the total number of description documents in the equipment defect feature dataset
  • is the number of datasets containing the information word ti.
  • the SVM Small Vector Machine
  • the SVM Small Vector Machine
  • a hyperplane can be found from the high-dimensional plane to serve as the dividing surface between the two categories, thereby ensuring the minimum classification error rate.
  • a device defect description feature is selected and input into the SVM classifier to perform defect classification prediction, and the features in the extracted equipment defect feature data set are compared and matched with the input defect description features, so as to perform classification prediction on each equipment defect description text and obtain the equipment defect grade prediction result TI.
  • other types of classifiers can also be used for defect classification prediction, which is not limited here.
  • accurate defect classification prediction can be achieved by calculating the importance of keywords in the equipment defect description features and combining them with the SVM classifier.
  • step S400 includes: step S402 , constructing a decision tree model, performing a correlation analysis on the equipment defect feature set and the equipment defect level prediction result according to the constructed decision tree model, and obtaining the correlation relationship between the defect features, defect influencing factors and component modules.
  • the equipment defect feature set G contains the defect feature information extracted from the equipment defect type, influencing factors, impact on equipment, impact on business, equipment manufacturer, defect area, alarm logic classification, defect component module, etc.
  • TI contains five defect level prediction results such as "very urgent", “urgent”, “major”, “general” and “other”
  • Y is the defect alarm root information label. Since the information contained in the equipment defect feature set G is discrete and independent of each other, the feature information can be digitized based on One-Hot encoding.
  • the quantized and encoded equipment defect feature set G is introduced into the association rule mining algorithm, and the association relationship between the features in the feature set G, i.e., the confidence, is calculated by formula (8) to form a formula such as Such a strong association.
  • Traverse the equipment defect feature set G in sequence set the minimum support and confidence in advance, prune the feature items below the minimum threshold, and then sort the candidate frequent item sets in descending order.
  • Create the root node of the tree structure continue to scan the equipment defect feature set G for the second time, process the feature items in each item set in order, and create a branch for each feature item set, that is, form a tree structure.
  • the root node of the tree structure is empty, recursively call the tree structure, continue to prune the items that do not meet the minimum threshold index, and determine whether a single path tree structure is finally formed. If so, list all feature item set combinations. If not, continue to call the tree structure until a single path tree structure is formed.
  • count( xa ⁇ xb ) represents the number of times the influencing factor xa and the equipment defect xb appear simultaneously; count( xa ) represents the number of times the influencing factor xa appears in the feature set.
  • the key to building a decision tree is to select the optimal partitioning features in the feature category, and to achieve as high a "purity" of the nodes as possible.
  • Information entropy and Gini index are used to measure the purity of defect alarms. Assuming that the proportion of samples in the kth category in the equipment defect alarm data set S is pk, the Gini value of S is:
  • the extracted discrete feature f may have multiple values. If there are N different values, N branch nodes will be generated when S is divided by f. Sn represents the nth branch point in the equipment defect data set S that takes the value of fn on the extracted feature f.
  • the Gini index obtained is expressed as:
  • the multidimensional data set ⁇ S, f ⁇ of the equipment defect alarm feature is input to build a decision tree model.
  • the value of Gini(S, f) is calculated by formula (10). If the calculated result satisfies formula (11), ⁇ S, f ⁇ is partitioned and stored until all partitions are processed. In this way, the support and confidence of the defect database are calculated to form a strong correlation between defect features, defect influencing factors and component modules.
  • the specific verification process can be: create training score and test score scoring functions, set the initial best score to 0, create two libraries to store the currently calculated score and the best score, and if the calculated score is less than the current best score, set this score to the best score.
  • the calculated accuracy is:
  • Step 1 Taking valve-based electronic equipment as an example, a physical model is established, and typical defects are selected by classifying the defect influencing factors. Local valve-based electronic equipment defect data is collected, and the defect characteristics and influencing factors are analyzed to generate a multi-dimensional influencing factor system to obtain the original equipment defect data set of valve-based electronic equipment.
  • Step 2 Preprocess the original data, determine the information gain of feature xi, construct the influencing factor function between equipment defects and factors, and construct the feature judgment function of the original equipment defect data set based on the information gain and influencing factor function.
  • Step 3 Evaluate the equipment defect features in the original equipment defect data set according to the feature evaluation function and filter out the initial equipment defect feature set.
  • Step 4 Construct a feature correlation function of the initial device defect feature set, and determine the correlation between each pair of features in the initial device defect feature set based on the feature correlation function.
  • Step 5 Cluster the initial equipment defect feature set according to the correlation between each pair of features in the initial equipment defect feature set to obtain a clustering result.
  • Step 6 Extract the equipment defect features from the clustering results to obtain the equipment defect feature set.
  • Step 7 Perform word segmentation processing on the equipment defect description features in the equipment defect feature set to obtain the word segmentation results.
  • Step 8 Determine the importance of keywords in the word segmentation results.
  • Step 9 Based on the importance of keywords, support vector machine is used to predict the defect grade of the equipment defect feature set to obtain the equipment defect grade prediction result.
  • Step 10 Integrate the equipment defect level prediction results and the equipment defect feature set to build a decision tree model. Perform a correlation analysis on the equipment defect feature set and the equipment defect level prediction results based on the constructed decision tree model to obtain the correlation between defect features, defect influencing factors and component modules.
  • the operation and maintenance personnel collected the defect data of the valve base electronic equipment of the DC transmission system in a certain area for example simulation, and used the feature correlation function in combination with the MapReduce framework to build a decision tree model to perform correlation analysis on the database.
  • valve base electronic equipment In actual applications, by analyzing the performance and operating parameters of valve base electronic equipment, it is found that the influencing factors include heating of optical fiber connectors, interruption of pigtails, transient interference of communication, software problems, equipment operation time, internal communication interruption, improper human operation, external temperature changes, etc.
  • the defective fault components of valve base electronic equipment that cause the VBE (VBScript Encoded Script, basic input and output system expansion bus) system to lose redundancy and do not temporarily affect DC operation include optical transmitter board, optical receiver board, CPU board, VBE system power supply, CLC (Central Logic Controller) interface board, VBE cabinet cooling system, signal relay and other components.
  • the validity analysis of the collected original data of the valve-based electronic equipment of the DC transmission system was carried out.
  • the influencing factors mainly include 9 categories, namely, fiber connector heating, pigtail interruption, instantaneous communication interference, software problems, equipment running time, internal communication interruption, improper human operation, and external temperature changes, which are numbered 1-9.
  • the faulty components mainly include 7 parts: optical transmitter board, optical receiver board, CPU board, VBE system power supply, CLC interface board, VBE cabinet cooling system, and signal relay, which are numbered ⁇ - ⁇ . After quantization coding, 900 valid equipment defect sets were obtained.
  • the embodiment of the present application also provides a method for analyzing the correlation of equipment defects, which can be applied in an application environment as shown in FIG. 1.
  • the terminal 102 communicates with the server 104 through the network.
  • the data storage system can store the data that the server 104 needs to process.
  • the data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers.
  • the operation and maintenance personnel can upload the collected equipment defect data to the server 104 through the terminal 102, and send the equipment defect analysis message to the server 104.
  • the server 104 responds to the message, obtains the equipment defect data, and matches the equipment defect data in the constructed equipment defect data rule base to obtain the equipment defect analysis result.
  • the equipment defect analysis result includes defect influencing factors, faulty component modules, and defect level prediction results, wherein the equipment defect data rule base is constructed using the above-mentioned equipment defect data rule base construction method.
  • the terminal 102 can be, but is not limited to, a personal computer, a laptop computer, a smart phone, a tablet computer, an Internet of Things device, and a portable wearable device.
  • the Internet of Things device can be a smart speaker, a smart TV, a smart air conditioner, a smart car device, etc.
  • the portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc.
  • the server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.
  • a method for analyzing the correlation between device defects is provided, and the method is described by taking the application of the method to the server 104 in FIG. 1 as an example, and includes the following steps:
  • Step S600 obtaining equipment defect data.
  • equipment defect data includes defect type characteristics, influencing factor characteristics, location characteristics, and component module characteristics.
  • equipment defect data can be obtained by operation and maintenance personnel collecting local equipment defect data.
  • Step S700 matching the equipment defect data in the constructed equipment defect data rule base to obtain equipment defect analysis results, wherein the equipment defect data rule base is constructed using the above-mentioned equipment defect data rule base construction method, and the equipment defect analysis results include defect influencing factors, faulty component modules and defect level prediction results.
  • the server can extract features from the equipment defect data, compare and match the extracted features in the established equipment defect data rule base, and obtain the equipment defect analysis results.
  • the extracted features include defect type A, defect level "very urgent", and defect influencing factor B. Then, because the defect level is "very urgent", among the numerous equipment defect data, the priority of this equipment defect data should be the highest.
  • the defect type A and defect influencing factor B are compared and matched in the equipment defect data rule base to match the association relationship related to defect type A and defect influencing factor B, that is, the strong rule.
  • the matched association relationship can be pushed to the operation and maintenance personnel as the equipment defect analysis result, so that the operation and maintenance personnel can check the component modules and the involved influencing factors one by one according to the defect type characteristics, defect influencing factors and component modules in the equipment defect analysis results, and eliminate the defects in a timely manner.
  • equipment defect data is obtained, and the equipment defect data is matched in the established equipment defect data rule base to obtain equipment defect analysis results, which include defect influencing factors, fault component modules, and defect level prediction results.
  • equipment defect analysis results include defect influencing factors, fault component modules, and defect level prediction results.
  • the device defect data includes an initial defect level
  • Step S700 includes:
  • Step S702 sorting the equipment defect data according to the initial defect level.
  • Step S704 extracting feature data of the sorted equipment defect data.
  • Step S706 keyword matching is performed on the feature data in the equipment defect data rule library in turn to match the corresponding equipment defect rule data, where the equipment defect rule data includes the confidence level of the equipment defect rule.
  • Step S708 sorting the equipment defect rule data according to the confidence level to obtain the equipment defect analysis result.
  • the initial defect level is the preliminary defect level judgment made by the operation and maintenance personnel according to the actual defect situation.
  • the initial defect level includes five defect level prediction results, such as "very urgent", “urgent”, “major”, “general” and “other".
  • the equipment defect data can be sorted in descending order according to the initial defect level, "very urgent", “urgent”, “major”, “general” and “other”.
  • the initial defect level of "very urgent” has the highest priority and the first order, which means that the equipment defect data with high priority should be matched first.
  • the characteristic data of the equipment defect data such as defect type characteristics, influencing factor characteristics and defect level characteristics, can be extracted.
  • the characteristic data extracted from each equipment defect data is matched with keywords in the equipment defect data rule library in turn, and the corresponding equipment defect rule data is matched. Since each equipment defect data corresponds to a corresponding confidence level, the equipment defect rule data can be sorted according to the confidence level, that is, sorted in descending order according to the confidence level. Further, the influencing factors with high confidence levels and their faulty component modules are matched first to obtain the equipment defect analysis results. It is understandable that the sorting method mentioned in the above embodiment can also be ascending sorting or other sorting methods, which are not limited here.
  • the specific order and process of comparison and matching are determined by the initial defect level and confidence, so that the defect analysis results with the highest accuracy can be arranged at the front, which can improve the accuracy of fault location and diagnosis, so that the operation and maintenance personnel can find the cause of the fault as quickly as possible and take corresponding measures to ensure the normal operation of the system and improve the energy utilization of the system.
  • the method further includes: receiving equipment defect analysis result feedback data, and optimizing the equipment defect data rule base according to the equipment defect analysis result feedback data.
  • the feedback data of equipment defect analysis results include feedback data obtained by operation and maintenance personnel based on the results of equipment defect analysis and taking corresponding measures to eliminate defects, based on whether the defects are completely eliminated, including the number of specific equipment defect rule data, fields with errors such as defect influencing factors and component modules, and the correct influencing factors-equipment defects-component modules.
  • the constructed equipment defect data rule base may not accurately and comprehensively cover all types of equipment defects. Therefore, it is necessary to optimize the equipment defect data rule base to improve the accuracy of defect data matching association.
  • the optimization training of the equipment defect data rule base based on the bird swarm intelligent optimization algorithm can be: divide the equipment defect rule data into a training set and a test set, initialize the bird swarm with chaos, input weights and biases, calculate the regression error fitness value according to the parameters of each bird, and record the optimal fitness value and weight bias. Continuously calculate and update the fitness value, and optimize the corresponding parameters at the same time to improve the accuracy of the decision tree model.
  • the matching accuracy of the equipment defect data rule base can be continuously improved to improve the accuracy of fault location and fault diagnosis.
  • steps in the flow chart involved in the above-mentioned embodiment are shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a part of the steps in the flow chart involved in the above-mentioned embodiment can include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
  • a person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through computer-readable instructions.
  • the computer-readable instructions can be stored in a non-volatile computer-readable storage medium. When the computer-readable instructions are executed, they can include the processes of the embodiments of the above-mentioned methods.
  • a device for constructing a rule base for equipment defect data including: a data acquisition module 810, a data extraction module 820, a defect grading module 830, a correlation analysis module 840 and a rule base construction module 850, wherein:
  • the data acquisition module 810 is used to acquire the original equipment defect data set
  • a data extraction module 820 configured to extract device defect features from the original device defect data set to obtain a device defect feature set, wherein the device defect feature set includes defect features, defect influencing factors, and component modules;
  • the defect grading module 830 is used to perform defect grading prediction on the equipment defect feature set to obtain an equipment defect grade prediction result;
  • a correlation analysis module 840 is used to perform correlation analysis on the device defect feature set and the device defect level prediction result to obtain the correlation relationship between defect features, defect influencing factors and component modules;
  • the rule base construction module 850 is used to construct the equipment defect data rule base based on the association relationship.
  • the above-mentioned device for constructing a rule base for equipment defect data extracts the equipment defect features of the original equipment defect data set to obtain an equipment defect feature set, performs defect grading prediction on the equipment defect feature set to obtain an equipment defect grade prediction result, and then performs correlation analysis on the equipment defect feature set and the equipment defect grade prediction result to obtain the defect features, defect influencing factors, and the correlation between component modules, and finally, constructs an equipment defect data rule base based on the correlation relationship.
  • the entire process, through the equipment defect data rule base can provide an accurate basis for the subsequent correlation analysis of equipment defects, fault tracing, and fault diagnosis, thereby improving the accuracy of fault diagnosis.
  • the data extraction module 820 is also used to construct a feature evaluation function of the original equipment defect data set, evaluate the equipment defect features in the original equipment defect data set according to the feature evaluation function, screen out the initial equipment defect feature set, construct a feature correlation function of the initial equipment defect feature set, and determine the correlation between each pair of features in the initial equipment defect feature set according to the feature correlation function, cluster the initial equipment defect feature set according to the correlation between each pair of features in the initial equipment defect feature set to obtain a clustering result, extract the equipment defect features in the clustering result, and obtain the equipment defect feature set.
  • the data extraction module 820 is also used to determine the information gain of the device defect features in the original device defect data set, construct an influencing factor function of the device defect features and the influencing factors, and construct a feature evaluation function of the original device defect data set based on the information gain and the influencing factor function; and according to the information gain of the device defect features, screen out the target device defect features, and add the target device defect features to the initial device defect feature set to update the initial device defect feature set.
  • the defect grading module 830 is further used to perform word segmentation processing on the device defect description features in the device defect feature set to obtain word segmentation results, and determine the importance of keywords in the word segmentation results;
  • support vector machine is used to predict the defect grade of equipment defect feature set and obtain the equipment defect grade prediction result.
  • a device for analyzing the correlation between equipment defects including: a data acquisition module 910 and a defect analysis module 920 , wherein:
  • the data acquisition module 910 is used to acquire equipment defect data.
  • the defect analysis module 920 is used to match the equipment defect data in the constructed equipment defect data rule library to obtain equipment defect analysis results.
  • the equipment defect data rule base is constructed using the above-mentioned equipment defect data rule base construction method, and the equipment defect analysis results include defect influencing factors, faulty component modules and defect level prediction results.
  • the above-mentioned device for analyzing the correlation of equipment defects obtains equipment defect data, matches the equipment defect data in the established equipment defect data rule base, and obtains equipment defect analysis results, which include defect influencing factors, faulty component modules, and defect level prediction results.
  • equipment defect analysis results which include defect influencing factors, faulty component modules, and defect level prediction results.
  • the defect analysis module 920 is also used to sort the equipment defect data according to the initial defect level, extract feature data of the sorted equipment defect data, and perform keyword matching on the feature data in the equipment defect data rule library in turn to match the corresponding equipment defect rule data.
  • the equipment defect rule data includes the confidence of the equipment defect rule. According to the confidence, the equipment defect rule data is sorted to obtain the equipment defect analysis result.
  • Each module in the above-mentioned equipment defect data rule base construction device and equipment defect correlation analysis device can be implemented in whole or in part by software, hardware and their combination.
  • the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
  • a computer device which may be a server, and its internal structure diagram may be shown in FIG10.
  • the computer device includes a processor, a memory, an input/output interface (Input/Output, referred to as I/O) and a communication interface.
  • the processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface.
  • the processor of the computer device is used to provide computing and control capabilities.
  • the memory of the computer device includes a non-volatile storage medium and an internal memory.
  • the non-volatile storage medium stores an operating system, a computer-readable instruction and a database.
  • the internal memory provides an environment for the operation of the operating system and the computer-readable instructions in the non-volatile storage medium.
  • the database of the computer device is used to store device defect data.
  • the input/output interface of the computer device is used to exchange information between the processor and an external device.
  • the communication interface of the computer device is used to communicate with an external terminal through a network connection.
  • FIG. 10 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.
  • the specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.
  • a computer device including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps in the above-mentioned device defect data rule base construction method or device defect correlation analysis method are implemented.
  • a computer-readable storage medium on which computer-readable instructions are stored.
  • the steps in the above-mentioned equipment defect data rule base construction method or equipment defect correlation analysis method are implemented.
  • a computer-readable instruction product including computer-readable instructions, which, when executed by a processor, implement the steps in the above-mentioned device defect data rule base construction method or device defect correlation analysis method.
  • the user information including but not limited to user device information, user personal information, etc.
  • data including but not limited to data used for analysis, stored data, displayed data, etc.
  • the user information and data involved in this application are all information and data authorized by the user or fully authorized by the party, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions.
  • any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory.
  • Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc.
  • Volatile memory can include random access memory (RAM) or external cache memory, etc.
  • RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • the database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database.
  • Non-relational databases may include distributed databases based on blockchain, etc., but are not limited to this.
  • the processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, etc., but are not limited to this.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Quality & Reliability (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

一种设备缺陷数据规则库构建方法、装置、计算机设备和存储介质。所述设备缺陷数据规则库构建方法包括：获取原始设备缺陷数据集，从原始设备缺陷数据集中国提取出设备缺陷特征集，对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果，对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系，基于关联关系构建设备缺陷数据规则库。本申请还提供了一种设备缺陷关联性分析方法、装置、计算机设备和存储介质，设备缺陷关联性分析方法包括：获取设备缺陷数据，将设备缺陷数据在设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果。采用本方法能提高故障诊断结果的准确度。

Description

设备缺陷数据规则库构建方法及设备缺陷关联性分析方法
相关申请的交叉引用
本申请要求于2022年11月18日提交中国专利局，申请号为202211447799X，申请名称为“设备缺陷关联性分析方法、装置、计算机设备和存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及输电安全技术领域，特别是涉及一种设备缺陷数据规则库构建方法、装置、计算机设备和存储介质，以及一种设备缺陷关联性分析方法、装置、计算机设备和存储介质。
背景技术
换流阀控制设备作为直流输电系统的核心设备，其正常的开通、关断和状态监测对直流系统的稳定运行起着至关重要的作用。换流阀控制设备作为直流输电系统的核心设备，其正常的开通、关断和状态监测对直流系统的稳定运行起着至关重要的作用。
然而，发明人意识到，在直流输电工程中，阀基电子设备经常出现故障，故障类型和故障发生的原因也不尽相同。事实上，换流阀控制设备的故障诊断过程中，有些故障需要直流系统停运后才能进行处理，且大多数故障缺陷由运维人员进行人为消缺和记录，并未对故障发生原因作深入分析。
由此可见，现有的换流阀控制设备的故障诊断结果存在准确率较低的问题。
发明内容
根据本申请公开的各种实施例，提供一种设备缺陷数据规则库构建方法、装置、计算机设备和存储介质，以及一种设备缺陷关联性分析方法、装置、计算机设备和存储介质。
一种设备缺陷数据规则库构建方法，包括：
获取原始设备缺陷数据集；
提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得缺陷特征、缺陷影响因素以及元件模块之间的关联关系；及
基于关联关系构建设备缺陷数据规则库。
一种设备缺陷关联性分析方法，包括：
获取设备缺陷数据；
将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果；及
其中，设备缺陷数据规则库采用上述的设备缺陷数据规则库构建方法构建，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
一种设备缺陷数据规则库构建装置，包括：
数据获取模块，用于获取原始设备缺陷数据集；
数据提取模块，用于提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
缺陷定级模块，用于对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
关联性分析模块，用于对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系；及
规则库构建模块，用于基于关联关系构建设备缺陷数据规则库。
一种设备缺陷关联性分析装置，包括：
数据获取模块，用于获取设备缺陷数据；
缺陷分析模块，用于将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果；及
其中，设备缺陷数据规则库采用如上述设备缺陷数据规则库构建方法构建，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
一种计算机设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时，使得一个或多个处理器执行上述设备缺陷数据规则库构建方法或上述设备缺陷关联性分析方法中的步骤。
一个或多个存储有计算机可读指令的计算机存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述设备缺陷数据规则库构建方法或上述设备缺陷关联性分析方法中的步骤。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为一个实施例中设备缺陷数据规则库构建方法或设备缺陷关联性分析方法的应用环境图；
图2为一个实施例中设备缺陷数据规则库构建方法的流程示意图；
图3为一个实施例中提取设备缺陷特征集步骤的流程示意图；
图4为另一个实施例中提取设备缺陷特征集步骤的流程示意图；
图5为一个实施例中设备缺陷数据规则库构建方法的详细流程示意图；
图6为一个实施例中设备缺陷关联性分析方法的流程示意图；
图7为一个实施例中将设备缺陷数据匹配步骤的流程示意图；
图8为一个实施例中设备缺陷数据规则库构建装置的结构框图；
图9为一个实施例中设备缺陷关联性分析装置的结构框图；
图10为一个实施例中计算机设备的内部结构图。
具体实施方式
为了使本申请的技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
本申请实施例提供的设备缺陷数据规则库构建方法，可以应用于如图1所示的应用环境中。其中，终端102通过网络与服务器104进行通信。数据存储系统可以存储服务器104需要处理的数据。数据存储系统可以集成在服务器104上，也可以放在云上或其他网络服务器上。具体的，可以是运维人员通过终端102上传原始设备缺陷数据集至服务器104，并发送规则库构建消息至服务器104，服务器104响应该消息，获取原始设备缺陷数据集，提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块，对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果，对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系，基于关联关系构建设备缺陷数据规则库。 其中，终端102可以但不限于是种个人计算机、笔记本电脑、智能手机、平板电脑、物联网设备和便携式可穿戴设备，物联网设备可为智能音箱、智能电视、智能空调、智能车载设备等。便携式可穿戴设备可为智能手表、智能手环、头戴设备等。服务器104可以用独立的服务器或者是多个服务器组成的服务器集群来实现。
在一个实施例中，如图2所示，提供了一种设备缺陷数据规则库构建方法，以该方法应用于图1中的服务器为例进行说明，包括以下步骤：
步骤S100，获取原始设备缺陷数据集。
原始设备缺陷数据集即指终端上传的通过前期数据准备工作得到的设备的缺陷数据集。原始设备缺陷数据集包括设备的缺陷类型、人为判定的初始缺陷等级、影响因素、故障元件、故障描述特征、时间信息以及天气情况等数据。设备以阀基电子设备为例，在实际应用中，原始设备缺陷数据集可以是采用以下方式得到：构建阀基电子设备物理模型，根据直流输电系统原理，分析阀基电子设备个元件组成及功能，分析设备缺陷特征及其影响因素。然后，采集当地的阀基电子设备的缺陷数据，根据特征对设备多种缺陷告警及其影响因素等相关信息进行归纳分析，得到阀基电子设备缺陷多维度影响因素体系和两者之间映射关系，得到原始设备缺陷数据集。具体的，可以是建立阀基电子设备的物理模型，通过缺陷影响因素分类，选取其中典型缺陷，包括信号异常、光功率过载、远端误码过量、磁路异常、网络丢包严重、CPU(Central Processing Unit,中央处理器)频率/电压异常、冷却效率降低等。采集当地阀基电子设备缺陷数据，每一条数据作为一个样本，选取n个数据样本，对缺陷特征及其影响因素进行分析，包括光纤接头发热、尾纤中断、通信瞬时干扰、软件问题、设备运行时间、内部通信中断、人为操作不当、外界温度变化等多个影响因素，生成多维度影响因素体系，得到阀基电子设备缺陷的原始设备缺陷数据集。
步骤S200，提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块。
在实际应用中，原始设备缺陷数据集中数据量非常多，且数据是杂乱无章的。因此，在获取原始设备缺陷特征集后，可以是对原始设备缺陷数据集进行数据清洗等数据预处理，然后基于特征相关性函数对该数据集中的特征进行评估，结合MapReduce框架，并行提取出原始设备缺陷数据集中的特征，包括缺陷类型特征、缺陷影响因素以及元件模块。得到设备缺陷特征集。例如，阀基电子设备的典型缺陷类型包括信号异常、光功率过载、远端误码过量、磁路异常、网络丢包严重、CPU频率/电压异常以及冷却效率降低等。通过分析阀基电子设备性能和运行参数等信息，得到其缺陷影响因素包括光纤接头发热、尾纤中断、通信瞬时干扰、软件问题、设备运行时间以及内部通信中断等，元件模块包括光发射板、光接收板、以及信号继电器等元件。
步骤S300，对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
提取出设备缺陷特征集后，可以对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。具体的，可以是通过分类器对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。本实施例中，可将缺陷定级划分为5个等级，分别为“非常紧急”、“紧急”、“重大”、“一般”和“其他”。可以理解的是，在其他实施例中，还可以是按照其他方式进行缺陷等级。
步骤S400，对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系。
当得到设备缺陷特征集和设备缺陷等级预测结果之后，可以是对设备缺陷特征集进行量化编码后，再对设备缺陷类型、缺陷影响因素以及元件模块之间进行关联性分析。具体的，可以确定影响缺陷告警产生的特征，通过计算缺陷数据库的支持度和置信度，形成缺陷影响因素-缺陷类型特征-元件模块三者之间的关联关系，若关联关系中，三者之间的相关性很强即为强关联关系(以下可简称为强规则)，则表征当出现该类型的缺陷时，导致缺陷产生的原因很大程度上与关联关系中对应的缺陷影响因素有关，出现故障的元件也很有可能是关联关系中对应的元件模块。可以理解的是，关联关系的数量可以是很多条，关联关系可以包括一对多、多对一以及多对多的关系，每一条关联关系均对应有相应的支持度和置信度，以表征缺陷类型特征、缺陷影响因素和元件模块之间的相关程度。例如，缺陷影响因 素a，b，c和d可对应缺陷类型特征A，元件模块1，2，3和4可对应的缺陷类型特征A。
步骤S500，基于关联关系构建设备缺陷数据规则库。
承接上述实施例，当得到缺陷影响因素-缺陷类型特征-元件模块三者之间的关联关系之后，由于得到的关联关系的数量非常多，因此，为了便于数据查找匹配，可以将多条关联关系、置信度、缺陷等级以及整合在一起，构建设备缺陷数据规则库。具体的，可以是根据置信度整合关联关系，得到设备缺陷规则数据库。
上述设备缺陷数据规则库构建方法中，通过提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果，再对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系，最后，基于关联关系构建设备缺陷数据规则库。整个过程，通过设备缺陷数据规则库，能够为后续的设备缺陷的关联分析和故障溯源和故障诊断提供准确的依据，进而提高故障诊断的准确度。
如图3所示，在一个实施例中，步骤200包括：
步骤S202，构建原始设备缺陷数据集的特征评判函数。
步骤S204，根据特征评判函数对原始设备缺陷数据集中的设备缺陷特征进行评估，筛选出初始设备缺陷特征集。
步骤S206，构建初始设备缺陷特征集的特征相关性函数，并根据特征相关性函数确定初始设备缺陷特征集中两两特征之间的相关性。
步骤S208，根据初始设备缺陷特征集中两两特征之间的相关性，对初始设备缺陷特征集进行聚类，得到聚类结果。
步骤S210，提取聚类结果中的设备缺陷特征，得到设备缺陷特征集。
特征评判函数也可称为特征评估函数，用于评价特征之间的相关性。特征相关性函数为评估两两特征之间的相似程度的函数。在另一个实施例中，特征评判函数F(x)的构建方式可以是：基于信息论的设备缺陷特征策略，利用信息论的相关知识衡量缺陷特征-标签及设备缺陷-影响因素之间的影响程度，构建特征评判函数。然后，利用特征评判函数对原始设备缺陷数据集中的设备缺陷特征进行评估，将特征评判值F(i)最大值对应的特征添加至初始设备缺陷特征集F中，迭代执行上述步骤，直到F中特征的个数达到预先设定的阈值，剩余的次要的特征自行添加至特征集F’，对于特征集F’可不做处理，也可删除。然后，构建初始设备缺陷特征集F的特征相关性函数：
Figure PCTCN2022134353-appb-000001
其中，x a，x b为初始设备缺陷特征集F中任意两个不同的特征，I(x a；Y|x b)为在特征x b的条件下，特征x a与标签Y之间的相关性，即x b的存在对x a与Y之间相关性的影响程度。通过特征相关性函数，确定初始设备缺陷特征集中两两特征之间的相关性。C(x a,x b)的值越小，则表征特征x a与特征x b之间的相似程度越大，反之，则表征特征x a与特征x b之间的相似程度越小。若x a代表缺陷类型特征，而x b代表影响因素特征，则函数C(x a,x b)的值越小，表明此缺陷类型特征与影响因素特征相似程度越大。基于此，可根据两两特征之间的相关性，初始设备缺陷特征集进行聚类，将设备缺陷的相似特征聚集在相同簇中，I＝{I 1，I 2，……，I k}得到为聚类后返回的簇的集合。对于簇集合I，调用MapReduce框架，根据框架中节点的个数均匀分配簇，每个节点通过调用Map函数利用PCA(Principle Compoent Analysis，主元成分分析法)算法对簇中的特征进行提取，得到设备缺陷特征集G。本实施例中，通过特征评判函数筛选出初始设备缺陷特征集，通过特征相关性函数对初始设备缺陷特征集进行特征提取，能够使得原本零散混乱的特征变得紧密相关，特征之间建立了联系。
如图4所示，在一个实施例中，步骤S202包括：
步骤S222，确定原始设备缺陷数据集中设备缺陷特征的信息增益，构建设备缺陷特征与影响因素的影响因素函数，基于信息增益和影响因素函数，构建原始设备缺陷数据集的特征评判函数。
步骤S204还包括：步骤S224，根据特征评判函数对原始设备缺陷数据集中的设备缺陷特征进行评估，筛选出初始设备缺陷特征集，根据设备缺陷特征的信息增益，筛选出目标设备缺陷特征，并将 目标设备缺陷特征添加至初始设备缺陷特征集，以更新初始设备缺陷特征集。
具体实施时，特征评判函数的构建过程可以是：对原始设备缺陷数据集进行预处理，x i为原始设备缺陷数据集中的任一特征，包括缺陷类型特征、影响因素特征、地点特征、元件模块特征等，上述特征可统称为设备缺陷特征，Y是对应的类别标签(即缺陷的类别)，得到设备缺陷特征x i的信息增益G(x i；Y)如式(2)所示。
G(x i；Y)＝H(Y)-H(Y|x i)      (2)
式中，H(Y)为关于类别标签Y的信息熵，H(Y|x i)为关于设备缺陷特征x i和类别Y的条件熵。
然后，构建设备缺陷-缺陷影响因素之间的影响因素函数Q，如式(3)所示。其中，S为原始设备缺陷数据集，Y为类别标签，x j为数据集S中的特征元素。
Figure PCTCN2022134353-appb-000002
式中，I(x j；Y)为互信息，I(x j；Y|x i)为关于变量x j和类别标签Y的条件互信息。
接着，基于上述信息增益和影响因素函数，同时考虑到缺陷特征-标签及缺陷-影响因素之间的影程度，由此构建原始设备缺陷数据集的特征评判函数F(i)，具体可表示为：
F(i)＝Q(i)+CG(x i；Y)(0≤C≤1)  (4)
其中，C为函数G(x i；Y)的权重参数。
进一步的，可以使用Hadoop中默认的文件块策略，将原始数据集的特征空间划分成大小相同的文件块Block；接着，文件块Block作为输入数据，Mapper节点通过调用Map函数以键值对<key，value>的形式，统计出每个特征的信息增益(key为特征名称，value为对应特征的信息增益)，组合每个键值对，得到特征信息增益集合A；最后，根据特征对应的信息增益值对集合A中元素降序排列，移除集合A中排序较为靠后的特征，重新组合得到新的缺陷设备缺陷特征矩阵X*。再将X*中信息增益值最大的特征添加至初始设备缺陷特征集，并依次计算备选特征的特征评判函数F(i)值，将F(i)最大值对应的特征放入F中，迭代执行上述步骤，直到F中特征的个数达到事先设定的阈值，得到初始设备缺陷特征集。本实施例中，通过信息增益和影响因素函数构建特征评判函数，能够使得特征评判函数准确衡量缺陷特征-标签及设备缺陷-影响因素之间的影响程度，进而使得提取出的设备缺陷特征集能够更具代表性。
如图5所示，在一个实施例中，步骤S300包括：
步骤302，对设备缺陷特征集中的设备缺陷描述特征进行分词处理，得到分词结果，确定分词结果中关键词的重要度，基于关键词的重要度，采用支持向量机对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
在实际应用中，设备缺陷特征集包含有设备缺陷描述文档，该文档中包含设备缺陷描述特征。运维人员预先构建有自定义词典，本实施例中，对设备缺陷特征集进行缺陷定级可以是：基于自定义词典，对每一条设备缺陷描述特征进行分词处理，然后，采用TF-IDF算法对设备缺陷描述特征中分词后的每个关键词进行处理，确定分词结果中关键词的重要度。具体的，TF-IDF的计算公式为：
TI ij＝tf i,j×idf i     (5)
其中，
Figure PCTCN2022134353-appb-000003
Figure PCTCN2022134353-appb-000004
式中，tf ij为词频，n i,j为描述词语ti在缺陷描述文档dj中出现的次数，∑ kn k,j为缺陷描述文档dj的所有描述词语数量之和。idf i为逆文档频率，|D|为设备缺陷特征数据集中的描述文档总数，|j:t i∈d j|为包含信息词语ti的数据集数量。
确定关键词的重要度之后，可以是采用SVM(Support Vector Machine,支持向量机)分类器对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。具体的，可以从高维平面中寻找一个超平面，以作为两类的分割面，从而保证最小的分类错误率。然后，选取一条设备缺陷描述特征输入至SVM 分类器，进行缺陷定级预测，将提取的设备缺陷特征数据集中的特征与输入的缺陷描述特征进行对比匹配，从而对每一条设备缺陷描述文本进行定级预测，得到设备缺陷等级预测结果TI。可以理解的是，在其他实施例中，还可以采用其他类型的分类器进行缺陷定级预测，在此不做限定。本实施例中，通过计算设备缺陷描述特征中关键词的重要度，结合SVM分类器，能够实现准确的缺陷定级预测。
如图5所示，在一个实施例中，步骤S400包括：步骤S402，构建决策树模型，根据构建的决策树模型对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系。
具体实施时，可以是融合设备缺陷特征集G、缺陷等级预测结果TI和类别标签Y，得到缺陷告警数据集S，并对缺陷告警数据集S进行归一化处理，得到S＝{G，TI，Y}。其中，设备缺陷特征集G包含了设备缺陷类型、影响因素、对设备影响、对业务影响、设备厂家、缺陷地区、告警逻辑分类、缺陷元件模块等提取的缺陷特征信息，而TI包含了“非常紧急”、“紧急”、“重大”、“一般”和“其他”等五种缺陷等级预测结果，Y则是缺陷告警根源信息标签。由于设备缺陷特征集G里包含的信息是离散且相互独立的，因此，可基于One-Hot编码对特征信息进行数字化处理。
将量化编码后的设备缺陷特征集G导入关联规则挖掘算法，通过式(8)计算特征集G里个特征之间的关联关系，即置信度，形成诸如
Figure PCTCN2022134353-appb-000005
这样的强关联。依次遍历设备缺陷特征集G，并预先设置最小支持度和置信度，剪枝低于最小阈值的特征项，然后将候选频繁项集按降序排列。创建树结构的根节点，继续二次扫描设备缺陷特征集G，对每个项集中的特征项按次序处理，并对每个特征项集创建一个分支，即形成树结构。树结构的根节点为空，递归调用树结构，继续剪枝不满足最小阈值指标的项，并判断最终是否形成单一路径的树结构，若是，则列举所有特征项集组合，若不是，则继续调用树结构，直至形成单一路径的树结构。
Figure PCTCN2022134353-appb-000006
式中，count(x a∪x b)表示影响因素x a与设备缺陷x b同时出现的次数；count(x a)表示影响因素x a在特征集中出现的次数。
构建决策树的关键是在特征类别中选择最优划分特征，需尽可能实现节点的高“纯度”。使用信息熵和基尼指数度量缺陷告警纯度。假设设备缺陷告警数据集S中第k类样本所占比例为pk，S的基尼值为：
Figure PCTCN2022134353-appb-000007
与信息熵相同，Gini(S)的值越小，代表设备缺陷告警数据集S的纯度越高。针对提取的离散特征f可能有多个取值，若有N个不同的取值时，则用f对S进行划分时会产生N个分支节点，用Sn表示包含在设备缺陷数据集S中全部在提取特征f上取值为fn的第n个分支点，以此得到的基尼指数表示为：
Figure PCTCN2022134353-appb-000008
选择基尼指数最小的值作为最优化分特征，即
f*＝arg f∈Fmin G ini(S，f)     (11)
然后，输入设备缺陷告警特征多维化数据集{S，f}，构建决策树模型。基于离散特征变量f在缺陷数据集S中的属性，通过式(10)计算Gini(S，f)的值，若计算出的结果满足式(11)，则将{S，f}进行分区存储，直到处理完所有分区，以此方式，计算缺陷数据库的支持度和置信度，形成缺陷特征-缺陷影响因素-元件模块三者之间的强关联关系。
进一步的，构建决策树模型后，需要对决策树模型的准确率进行验证。具体的验证过程可以是：创建训练分数和测试分数评分功能，设置初始最佳分数为0，创建两个库用于存储当前计算得到的分数和最佳分数，如果计算得到的分数小于当前最佳分数，则将此分数设置为最佳分数。用决策树性能指标分类正确率作为决策树分类器对于设备缺陷特征的评价指标，计算的准确率为：
Figure PCTCN2022134353-appb-000009
式中，
Figure PCTCN2022134353-appb-000010
表示参数为真则等于1，否则等于0。
Figure PCTCN2022134353-appb-000011
分别表示第i个设备缺陷告警数据的真实值和决策树分类值。
本实施例中，通过构建决策树模型对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，能够得到准确的缺陷特征-缺陷影响因素-元件模块三者之间的关联关系。
为了对本申请提供的设备缺陷数据规则库构建方法做出清楚的说明，下面结合一个详细实施例进行说明，该实施例包括以下内容：
步骤1：以阀基电子设备为例，建立物理模型，通过缺陷影响因素分类，选取其中典型缺陷。采集当地阀基电子设备缺陷数据，对缺陷特征及其影响因素进行分析，生成多维度影响因素体系，得到阀基电子设备的原始设备缺陷数据集。
步骤2：对原始数据进行预处理，确定特征xi的信息增益，构建设备缺陷-因素之间的影响因素函数，根据信息增益和影响因素函数，构建原始设备缺陷数据集的特征评判函数。
步骤3：根据特征评判函数对原始设备缺陷数据集中的设备缺陷特征进行评估，筛选出初始设备缺陷特征集。
步骤4：构建初始设备缺陷特征集的特征相关性函数，并根据特征相关性函数确定初始设备缺陷特征集中两两特征之间的相关性。
步骤5：根据初始设备缺陷特征集中两两特征之间的相关性，对初始设备缺陷特征集进行聚类，得到聚类结果。
步骤6：提取聚类结果中的设备缺陷特征，得到设备缺陷特征集。
步骤7：对设备缺陷特征集中的设备缺陷描述特征进行分词处理，得到分词结果。
步骤8：确定分词结果中关键词的重要度。
步骤9：基于关键词的重要度，采用支持向量机对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
步骤10：融合设备缺陷等级预测结果和设备缺陷特征集，构建决策树模型，根据构建的决策树模型对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系。
运维人员采集了某地区的直流输电系统阀基电子设备缺陷数据进行算例仿真，利用特征相关性函数，结合MapReduce框架，构建决策树模型对数据库进行关联性分析。
在实际应用中，通过分析阀基电子设备性能和运行参数等信息，得到其影响因素包括光纤接头发热、尾纤中断、通信瞬时干扰、软件问题、设备运行时间、内部通信中断、人为操作不当、外界温度变化等。在导致VBE(VBScript Encoded Script，基本输入输出系统扩展总线)系统失去冗余和暂不影响直流运行的阀基电子设备缺陷故障元件中，包含了光发射板、光接收板、CPU板卡、VBE系统电源、CLC(Central Logic Controller，中央逻辑控制器)接口板卡、VBE屏柜冷却系统、信号继电器等元件。
对采集的直流输电系统阀基电子设备原始数据进行有效性分析。分类归纳缺陷类型主要有8类，包括信号异常、光功率过载、远端误码过量、磁路异常、网络丢包严重、ESD损坏、CPU频率/电压异常、冷却效率降低，使用A-H进行编号；影响因素则主要包含9类，分别为光纤接头发热、尾纤中断、通信瞬时干扰、软件问题、设备运行时间、内部通信中断、人为操作不当、外界温度变化，使用1-9进行编号。故障元件主要包括光发射板、光接收板、CPU板卡、VBE系统电源、CLC接口板卡、VBE屏柜冷却系统、信号继电器7部分，用α-η进行编号。量化编码后得到900条有效设备缺陷集。
通过将整合编号后的设备缺陷数据库通过决策树算法进行分析，得到相应的强规则结果，根据最小阈值条件，并得到加入特征的正确率结果如表1所示。
在一个实施例中，本申请实施例还提供了一种设备缺陷关联性分析方法，该方法可以应用于如图1所示的应用环境中。其中，终端102通过网络与服务器104进行通信。数据存储系统可以存储服务器104需要处理的数据。数据存储系统可以集成在服务器104上，也可以放在云上或其他网络服务器上。 具体的，可以是运维人员通过终端102上传采集的设备缺陷数据至服务器104，并发送设备缺陷分析消息至服务器104，服务器104响应该消息，获取设备缺陷数据，将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果，其中，设备缺陷数据规则库采用上述的设备缺陷数据规则库构建方法构建。其中，终端102可以但不限于是种个人计算机、笔记本电脑、智能手机、平板电脑、物联网设备和便携式可穿戴设备，物联网设备可为智能音箱、智能电视、智能空调、智能车载设备等。便携式可穿戴设备可为智能手表、智能手环、头戴设备等。服务器104可以用独立的服务器或者是多个服务器组成的服务器集群来实现。
表1直流输电系统阀基电子设备缺陷数据强规则结果及加入特征的正确率
Figure PCTCN2022134353-appb-000012
在一个实施例中，如图6所示，提供了一种设备缺陷关联性分析方法，以该方法应用于图1中的服务器104为例进行说明，包括以下步骤：
步骤S600，获取设备缺陷数据。
在实际应用中，设备缺陷数据包括缺陷类型特征、影响因素特征、地点特征以及元件模块特征等数据。同样的，设备缺陷数据可以是由运维人员采集当地的设备的缺陷数据得到。设备缺陷数据可以是多条，每一条数据均包括缺陷类型特征、影响因素特征、地点特征以及元件模块特征等数据。
步骤S700，将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果，其中，设备缺陷数据规则库采用上述的设备缺陷数据规则库构建方法构建，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
服务器在得到设备缺陷数据后，可以是提取设备缺陷数据中的特征，将提取的特征在已构建的设备缺陷数据规则库中进行对比匹配，得到设备缺陷分析结果。例如，针对一条设备缺陷数据，提取的特征包括缺陷类型A，缺陷等级为“非常紧急”，缺陷影响因素B，然后，由于缺陷等级为“非常紧急”，因此，在众多的设备缺陷数据中，此设备缺陷数据的优先级应最大，将缺陷类型A，缺陷影响因素B在设备缺陷数据规则库中进行对比匹配，匹配出与缺陷类型A和缺陷影响因素B相关的关联关系即强规则。然后，可以将匹配出的关联关系作为设备缺陷分析结果，推送给运维人员，以便运维人员根据设备缺陷分析结果中的缺陷类型特征、缺陷影响因素以及元件模块，逐一检查元件模块和涉及的影响因素，以及时进行消缺。
上述设备缺陷关联性分析方法中，获取设备缺陷数据，将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。整个过程，通过将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，便能实现自动化的设备缺陷的关联分析和故障定位诊断，提高了故障诊断的准确度。
如图7所示，在一个实施例中，设备缺陷数据包括初始缺陷等级；
步骤S700包括：
步骤S702，根据初始缺陷等级，对设备缺陷数据进行排序。
步骤S704，提取排序后的设备缺陷数据的特征数据。
步骤S706，依次将特征数据在设备缺陷数据规则库中进行关键词匹配，匹配出对应的设备缺陷规 则数据，设备缺陷规则数据包括设备缺陷规则的置信度。
步骤S708，根据置信度，对设备缺陷规则数据进行排序，得到设备缺陷分析结果。
初始缺陷等级为运维人员根据实际的缺陷情况做出的初步的缺陷等级判定，同样的，初始缺陷等级包含了“非常紧急”、“紧急”、“重大”、“一般”和“其他”等五种缺陷等级预测结果。具体实施时，可以是根据初始缺陷等级，按照“非常紧急”、“紧急”、“重大”、“一般”和“其他”对设备缺陷数据进行降序排序，初始缺陷等级为“非常紧急”的优先级最高，顺序最靠前，即意味着应最先匹配优先级高的设备缺陷数据。排序处理后，可提取设备缺陷数据的特征数据如缺陷类型特征、影响因素特征和缺陷等级特征等，然后，按照排序后的设备缺陷数据的顺序，依次将每一条设备缺陷数据提取出的特征数据在设备缺陷数据规则库中进行关键词匹配，匹配出对应的设备缺陷规则数据。由于每一条设备缺陷数据均对应有相应的置信度，因此，可以根据置信度，对设备缺陷规则数据进行排序，即按照置信度进行降序排序，进一步，优先匹配置信度高的影响因素及其故障元件模块，得到设备缺陷分析结果。可以理解的是，上述实施例中提到的排序方式还可以是升序排序或其他排序方式，在此不做限定。本实施例中，通过初始缺陷等级和置信度，确定对比匹配的具体顺序和过程，能够使得准确度最高的缺陷分析结果排列在最前面，能够提高故障定位和诊断的准确率，使得运维人员能够以最快的时间查找出故障原因，并采取相应的措施，保障系统的正常运行，提高系统的能量利用率。
在一个实施例中，步骤S700之后，还包括：接收设备缺陷分析结果反馈数据，根据设备缺陷分析结果反馈数据优化设备缺陷数据规则库。
设备缺陷分析结果反馈数据包括运维人员根据设备缺陷分析结果进行实地勘验，并采取相应的消除缺陷的措施后，根据缺陷是否完全消除得到的反馈数据，包括具体的设备缺陷规则数据的编号、存在误差的字段如缺陷影响因素以及元件模块、以及正确的影响因素-设备缺陷-元件模块的关联关系等数据。由于实际的设备缺陷情况可能存在错综复杂的情况，构建的设备缺陷数据规则库可能无法准确全面覆盖到所有类型的设备缺陷，因此，需要对设备缺陷数据规则库进行优化，提高缺陷数据匹配关联的准确性。具体的，可以是构建基于鸟群优化算法的设备缺陷数据融合模型，对新的缺陷数据和生成的设备缺陷数据规则库进行匹配，根据回归误差适应度值，更新记录最有适应度和对应的参数，同时，接收设备缺陷分析结果反馈数据，进一步对设备缺陷规则数据的置信度进行调整，不断提高缺陷特征-缺陷影响因素-元件模块之间匹配预测的准确性和效率。例如，若运维人员反馈匹配出的设备缺陷规则数据A中缺陷影响因素存在误差，或者元件模块有误等情况，则根据实际情况与匹配出的结果的误差，相应下调该设备缺陷规则数据A的置信度。
具体实施时，基于鸟群智能优化算法对设备缺陷数据规则库进行优化训练可以是：将设备缺陷规则数据分为训练集和测试集，混沌初始化鸟群，输入权重和偏置构成，根据每只鸟的参数计算回归误差适应度值，并记录最优的适应度值和权重偏置。不断计算更新适应度值，同时优化对应的参数，以提高决策树模型的正确率。本实施例中，通过设备缺陷数据规则库进行不断训练，能够不断提高设备缺陷数据规则库的匹配准确度，以提高故障定位和故障诊断的准确度。
应该理解的是，虽然如上所述的实施例所涉及的流程图中的个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，如上所述的实施例所涉及的流程图中的至少一部分步骤可以包括多个步骤或者多个阶段，这些步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤中的步骤或者阶段的至少一部分轮流或者交替地执行。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。
在一个实施例中，如图8所示，提供了一种设备缺陷数据规则库构建装置，包括：数据获取模块810、数据提取模块820、缺陷定级模块830、关联性分析模块840和规则库构建模块850，其中：
数据获取模块810，用于获取原始设备缺陷数据集；
数据提取模块820，用于提取所述原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，所述设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
缺陷定级模块830，用于对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
关联性分析模块840，用于对所述设备缺陷特征集和所述设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系；
规则库构建模块850，用于基于所述关联关系构建所述设备缺陷数据规则库。
上述设备缺陷数据规则库构建装置，通过提取原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果，再对设备缺陷特征集和设备缺陷等级预测结果进行关联性分析，得到缺陷特征、缺陷影响因素以及元件模块之间的关联关系，最后，基于关联关系构建设备缺陷数据规则库。整个过程，通过设备缺陷数据规则库，能够为后续的设备缺陷的关联分析和故障溯源和故障诊断提供准确的依据，进而提高故障诊断的准确度。
在一个实施例中，数据提取模块820还用于构建原始设备缺陷数据集的特征评判函数，根据特征评判函数对原始设备缺陷数据集中的设备缺陷特征进行评估，筛选出初始设备缺陷特征集，构建初始设备缺陷特征集的特征相关性函数，并根据特征相关性函数确定初始设备缺陷特征集中两两特征之间的相关性，根据初始设备缺陷特征集中两两特征之间的相关性，对初始设备缺陷特征集进行聚类，得到聚类结果，提取聚类结果中的设备缺陷特征，得到设备缺陷特征集。
在一个实施例中，数据提取模块820还用于确定原始设备缺陷数据集中设备缺陷特征的信息增益，构建设备缺陷特征与影响因素的影响因素函数，基于信息增益和影响因素函数，构建原始设备缺陷数据集的特征评判函数；以及根据设备缺陷特征的信息增益，筛选出目标设备缺陷特征，并将目标设备缺陷特征添加至初始设备缺陷特征集，以更新初始设备缺陷特征集。
在一个实施例中，缺陷定级模块830还用于对设备缺陷特征集中的设备缺陷描述特征进行分词处理，得到分词结果，确定分词结果中关键词的重要度；
基于关键词的重要度，采用支持向量机对设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
在一个实施例中，如图9所示，提供了一种设备缺陷关联性分析装置，包括：数据获取模块910和缺陷分析模块920，其中：
数据获取模块910，用于获取设备缺陷数据。
缺陷分析模块920，用于将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果。
其中，设备缺陷数据规则库采用上述设备缺陷数据规则库构建方法构建，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
上述设备缺陷关联性分析装置，获取设备缺陷数据，将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果，设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。整个过程，通过将设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，便能实现自动化的设备缺陷的关联分析和故障定位诊断，提高了故障诊断的准确度。
在一个实施例中，缺陷分析模块920还用于根据初始缺陷等级，对设备缺陷数据进行排序，提取排序后的设备缺陷数据的特征数据，依次将特征数据在设备缺陷数据规则库中进行关键词匹配，匹配出对应的设备缺陷规则数据，设备缺陷规则数据包括设备缺陷规则的置信度，根据置信度，对设备缺陷规则数据进行排序，得到设备缺陷分析结果。
上述设备缺陷数据规则库构建装置和设备缺陷关联性分析装置中的个模块可全部或部分通过软件、硬件及其组合来实现。上述模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上个模块对应的操作。
在一个实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构图可以如图10所示。该计算机设备包括处理器、存储器、输入/输出接口(Input/Output，简称I/O)和通信接口。其中，处理器、存储器和输入/输出接口通过系统总线连接，通信接口通过输入/输出接口连接到系统总线。 其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质和内存储器。该非易失性存储介质存储有操作系统、计算机可读指令和数据库。该内存储器为非易失性存储介质中的操作系统和计算机可读指令的运行提供环境。该计算机设备的数据库用于存储设备缺陷数据。该计算机设备的输入/输出接口用于处理器与外部设备之间交换信息。该计算机设备的通信接口用于与外部的终端通过网络连接通信。该计算机可读指令被处理器执行时以实现一种设备缺陷数据规则库构建方法和一种设备缺陷关联性分析方法。
本领域技术人员可以理解，图10中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一个实施例中，提供了一种计算机设备，包括存储器和处理器，存储器中存储有计算机可读指令，该处理器执行计算机可读指令时实现上述设备缺陷数据规则库构建方法或设备缺陷关联性分析方法中的步骤。
在一个实施例中，提供了一种计算机可读存储介质，其上存储有计算机可读指令，计算机可读指令被处理器执行时实现上述设备缺陷数据规则库构建方法或设备缺陷关联性分析方法中的步骤。
在一个实施例中，提供了一种计算机可读指令产品，包括计算机可读指令，该计算机可读指令被处理器执行时实现上述设备缺陷数据规则库构建方法或设备缺陷关联性分析方法中的步骤。
需要说明的是，本申请所涉及的用户信息(包括但不限于用户设备信息、用户个人信息等)和数据(包括但不限于用于分析的数据、存储的数据、展示的数据等)，均为经用户授权或者经过方充分授权的信息和数据，且相关数据的收集、使用和处理需要遵守相关国家和地区的相关法律法规和标准。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、数据库或其它介质的任何引用，均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory，ROM)、磁带、软盘、闪存、光存储器、高密度嵌入式非易失性存储器、阻变存储器(ReRAM)、磁变存储器(Magnetoresistive Random Access Memory，MRAM)、铁电存储器(Ferroelectric Random Access Memory，FRAM)、相变存储器(Phase Change Memory，PCM)、石墨烯存储器等。易失性存储器可包括随机存取存储器(Random Access Memory，RAM)或外部高速缓冲存储器等。作为说明而非局限，RAM可以是多种形式，比如静态随机存取存储器(Static Random Access Memory，SRAM)或动态随机存取存储器(Dynamic Random Access Memory，DRAM)等。本申请所提供的各实施例中所涉及的数据库可包括关系型数据库和非关系型数据库中至少一种。非关系型数据库可包括基于区块链的分布式数据库等，不限于此。本申请所提供的各实施例中所涉及的处理器可为通用处理器、中央处理器、图形处理器、数字信号处理器、可编程逻辑器、基于量子计算的数据处理逻辑器等，不限于此。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (18)

  1. 一种设备缺陷数据规则库构建方法，包括：
    获取原始设备缺陷数据集；
    提取所述原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，所述设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
    对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
    对所述设备缺陷特征集和所述设备缺陷等级预测结果进行关联性分析，得所述缺陷特征、所述缺陷影响因素以及所述元件模块之间的关联关系；及
    基于所述关联关系构建所述设备缺陷数据规则库。
  2. 根据权利要求1所述的设备缺陷数据规则库构建方法，其中，所述提取所述原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集包括：
    构建所述原始设备缺陷数据集的特征评判函数；
    根据所述特征评判函数对所述原始设备缺陷数据集中的特征进行评估，筛选出初始设备缺陷特征集；
    构建所述初始设备缺陷特征集的特征相关性函数，并根据所述特征相关性函数确定所述初始设备缺陷特征集中两两特征之间的相关性；
    根据所述初始设备缺陷特征集中两两特征之间的相关性，对所述初始设备缺陷特征集进行聚类，得到聚类结果；及
    提取所述聚类结果中的设备缺陷特征，得到设备缺陷特征集。
  3. 根据权利要求2所述的设备缺陷数据规则库构建方法，其中，所述构建所述原始设备缺陷数据集的特征评判函数包括：
    确定所述原始设备缺陷数据集中设备缺陷特征的信息增益；
    构建设备缺陷特征与影响因素的影响因素函数；
    基于所述信息增益和所述影响因素函数，构建所述原始设备缺陷数据集的特征评判函数；及
    所述方法还包括：
    根据所述各数据特征的信息增益，筛选出目标设备缺陷特征，并将所述目标设备缺陷特征添加至所述初始设备缺陷特征集，以更新所述初始设备缺陷特征集。
  4. 根据权利要求3所述的设备缺陷数据规则库构建方法，其中，所述确定所述原始设备缺陷数据集中设备缺陷特征的信息增益包括：
    获取所述原始设备缺陷数据集中类别标签的信息熵，以及设备缺陷特征与所述类别标签的条件熵；及
    根据所述信息熵和所述条件熵，确定原始设备缺陷数据集中设备缺陷特征的信息增益。
  5. 根据权利要求3所述的设备缺陷数据规则库构建方法，其中，所述构建设备缺陷特征与影响因素的影响因素函数包括：
    获取所述原始设备缺陷数据集中类别标签与设备缺陷特征的互信息和条件互信息；及
    根据所述互信息和所述条件互信息，构建设备缺陷特征与影响因素的影响因素函数。
  6. 根据权利要求1所述的设备缺陷数据规则库构建方法，其中，所述对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果包括：
    对所述设备缺陷特征集中的设备缺陷描述特征进行分词处理，得到分词结果；
    确定所述分词结果中关键词的重要度；及
    基于所述关键词的重要度，采用支持向量机对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
  7. 一种设备缺陷关联性分析方法，包括：
    获取设备缺陷数据；
    将所述设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果；及
    其中，所述设备缺陷数据规则库采用如权利要求1至6任意一项所述的设备缺陷数据规则库构建方法构建，所述设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
  8. 根据权利要求1所述的设备缺陷关联性分析方法，其中，所述设备缺陷数据包括初始缺陷等级；
    所述将所述设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果包括：
    根据所述初始缺陷等级，对所述设备缺陷数据进行排序；
    提取排序后的所述设备缺陷数据的特征数据；
    依次将所述特征数据在所述设备缺陷数据规则库中进行关键词匹配，匹配出对应的设备缺陷规则数据，所述设备缺陷规则数据包括设备缺陷规则的置信度；及
    根据所述置信度，对所述设备缺陷规则数据进行排序，得到设备缺陷分析结果。
  9. 一种设备缺陷数据规则库构建装置，包括：
    数据获取模块，用于获取原始设备缺陷数据集；
    数据提取模块，用于提取所述原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，所述设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
    缺陷定级模块，用于对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
    关联性分析模块，用于对所述设备缺陷特征集和所述设备缺陷等级预测结果进行关联性分析，得到所述缺陷特征、所述缺陷影响因素以及所述元件模块之间的关联关系；及
    规则库构建模块，用于基于所述关联关系构建所述设备缺陷数据规则库。
  10. 一种设备缺陷关联性分析装置，包括：
    数据获取模块，用于获取设备缺陷数据；
    缺陷分析模块，用于将所述设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果；及
    其中，所述设备缺陷数据规则库采用如权利要求1至6任意一项所述的设备缺陷数据规则库构建方法构建，所述设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
  11. 一种计算机设备，包括存储器及一个或多个处理器，所述存储器中储存有计算机可读指令，所述计算机可读指令被所述一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    获取原始设备缺陷数据集；
    提取所述原始设备缺陷数据集的设备缺陷特征，得到设备缺陷特征集，所述设备缺陷特征集包括缺陷特征、缺陷影响因素以及元件模块；
    对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果；
    对所述设备缺陷特征集和所述设备缺陷等级预测结果进行关联性分析，得所述缺陷特征、所述缺陷影响因素以及所述元件模块之间的关联关系；及
    基于所述关联关系构建所述设备缺陷数据规则库。
  12. 根据权利要求11所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    构建所述原始设备缺陷数据集的特征评判函数；
    根据所述特征评判函数对所述原始设备缺陷数据集中的特征进行评估，筛选出初始设备缺陷特征集；
    构建所述初始设备缺陷特征集的特征相关性函数，并根据所述特征相关性函数确定所述初始设备缺陷特征集中两两特征之间的相关性；
    根据所述初始设备缺陷特征集中两两特征之间的相关性，对所述初始设备缺陷特征集进行聚类，得到聚类结果；及
    提取所述聚类结果中的设备缺陷特征，得到设备缺陷特征集。
  13. 根据权利要求12所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    确定所述原始设备缺陷数据集中设备缺陷特征的信息增益；
    构建设备缺陷特征与影响因素的影响因素函数；
    基于所述信息增益和所述影响因素函数，构建所述原始设备缺陷数据集的特征评判函数；及
    根据所述各数据特征的信息增益，筛选出目标设备缺陷特征，并将所述目标设备缺陷特征添加至所述初始设备缺陷特征集，以更新所述初始设备缺陷特征集。
  14. 根据权利要求13所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    获取所述原始设备缺陷数据集中类别标签的信息熵，以及设备缺陷特征与所述类别标签的条件熵；及
    根据所述信息熵和所述条件熵，确定原始设备缺陷数据集中设备缺陷特征的信息增益。
  15. 根据权利要求13所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    获取所述原始设备缺陷数据集中类别标签与设备缺陷特征的互信息和条件互信息；及
    根据所述互信息和所述条件互信息，构建设备缺陷特征与影响因素的影响因素函数。
  16. 根据权利要求11所述的计算机设备，其中，所述处理器执行所述计算机可读指令时还执行以下步骤：
    对所述设备缺陷特征集中的设备缺陷描述特征进行分词处理，得到分词结果；
    确定所述分词结果中关键词的重要度；及
    基于所述关键词的重要度，采用支持向量机对所述设备缺陷特征集进行缺陷定级预测，得到设备缺陷等级预测结果。
  17. 一个或多个存储有计算机可读指令的计算机存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    获取设备缺陷数据；
    将所述设备缺陷数据在已构建的设备缺陷数据规则库中进行匹配，得到设备缺陷分析结果；及
    其中，所述设备缺陷数据规则库采用如权利要求1至6任意一项所述的设备缺陷数据规则库构建方法构建，所述设备缺陷分析结果包括缺陷影响因素、故障元件模块以及缺陷等级预测结果。
  18. 根据权利要求17所述的存储介质，其中，所述设备缺陷数据包括初始缺陷等级；
    所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    根据所述初始缺陷等级，对所述设备缺陷数据进行排序；
    提取排序后的所述设备缺陷数据的特征数据；
    依次将所述特征数据在所述设备缺陷数据规则库中进行关键词匹配，匹配出对应的设备缺陷规则数据，所述设备缺陷规则数据包括设备缺陷规则的置信度；及
    根据所述置信度，对所述设备缺陷规则数据进行排序，得到设备缺陷分析结果。
PCT/CN2022/134353 2022-11-18 2022-11-25 设备缺陷数据规则库构建方法及设备缺陷关联性分析方法 WO2024103436A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211447799.XA CN115934393A (zh) 2022-11-18 2022-11-18 设备缺陷关联性分析方法、装置、计算机设备和存储介质
CN202211447799.X 2022-11-18

Publications (1)

Publication Number Publication Date
WO2024103436A1 true WO2024103436A1 (zh) 2024-05-23

Family

ID=86650003

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/134353 WO2024103436A1 (zh) 2022-11-18 2022-11-25 设备缺陷数据规则库构建方法及设备缺陷关联性分析方法

Country Status (2)

Country Link
CN (1) CN115934393A (zh)
WO (1) WO2024103436A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118278827A (zh) * 2024-06-04 2024-07-02 临沂红阳管业有限公司 一种基于塑料管质量检测的管材生产设备管理方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130279794A1 (en) * 2012-04-19 2013-10-24 Applied Materials Israel Ltd. Integration of automatic and manual defect classification
CN112949874A (zh) * 2021-03-04 2021-06-11 国网江苏省电力有限公司南京供电分公司 一种配电终端缺陷特征自诊断方法及系统
CN113313409A (zh) * 2021-06-16 2021-08-27 中国南方电网有限责任公司 基于数据关联的电力系统二次设备缺陷分析方法及系统
CN114090647A (zh) * 2021-10-22 2022-02-25 国家电网公司西南分部 一种电力通信设备缺陷关联性分析方法及缺陷排查方法
CN114169406A (zh) * 2021-11-17 2022-03-11 西安理工大学 基于对称不确定性联合条件熵的特征选择方法
CN114860931A (zh) * 2022-04-26 2022-08-05 华北电力大学 一种基于Voting Classifier模型的继电保护缺陷文本定级方法
CN115329144A (zh) * 2022-08-09 2022-11-11 中国银行股份有限公司 一种产品缺陷的根因确定方法及装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130279794A1 (en) * 2012-04-19 2013-10-24 Applied Materials Israel Ltd. Integration of automatic and manual defect classification
CN112949874A (zh) * 2021-03-04 2021-06-11 国网江苏省电力有限公司南京供电分公司 一种配电终端缺陷特征自诊断方法及系统
CN113313409A (zh) * 2021-06-16 2021-08-27 中国南方电网有限责任公司 基于数据关联的电力系统二次设备缺陷分析方法及系统
CN114090647A (zh) * 2021-10-22 2022-02-25 国家电网公司西南分部 一种电力通信设备缺陷关联性分析方法及缺陷排查方法
CN114169406A (zh) * 2021-11-17 2022-03-11 西安理工大学 基于对称不确定性联合条件熵的特征选择方法
CN114860931A (zh) * 2022-04-26 2022-08-05 华北电力大学 一种基于Voting Classifier模型的继电保护缺陷文本定级方法
CN115329144A (zh) * 2022-08-09 2022-11-11 中国银行股份有限公司 一种产品缺陷的根因确定方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118278827A (zh) * 2024-06-04 2024-07-02 临沂红阳管业有限公司 一种基于塑料管质量检测的管材生产设备管理方法及系统

Also Published As

Publication number Publication date
CN115934393A (zh) 2023-04-07

Similar Documents

Publication Publication Date Title
US11113124B2 (en) Systems and methods for quickly searching datasets by indexing synthetic data generating models
WO2021042843A1 (zh) 告警信息的决策方法、装置、计算机设备及存储介质
CN111612041B (zh) 异常用户识别方法及装置、存储介质、电子设备
US8965893B2 (en) System and method for grouping multiple streams of data
WO2023226423A1 (zh) 一种芯片辅助设计方法、装置、设备及非易失性存储介质
US20140207786A1 (en) System and methods for computerized information governance of electronic documents
CN105471647B (zh) 一种电力通信网故障定位方法
CN113821657A (zh) 基于人工智能的图像处理模型训练方法及图像处理方法
CN111368867B (zh) 档案归类方法及系统、计算机可读存储介质
CN112306820B (zh) 一种日志运维根因分析方法、装置、电子设备及存储介质
CN113254255A (zh) 一种云平台日志的分析方法、系统、设备及介质
WO2024103436A1 (zh) 设备缺陷数据规则库构建方法及设备缺陷关联性分析方法
US20190050672A1 (en) INCREMENTAL AUTOMATIC UPDATE OF RANKED NEIGHBOR LISTS BASED ON k-th NEAREST NEIGHBORS
CN112685324A (zh) 一种生成测试方案的方法及系统
CN115619245A (zh) 一种基于数据降维方法的画像构建和分类方法及系统
CN117221087A (zh) 告警根因定位方法、装置及介质
CN115114126A (zh) 获取层级化数据结构以及处理日志条目方法和电子设备
CN116841779A (zh) 异常日志检测方法、装置、电子设备和可读存储介质
CN110544047A (zh) 一种不良数据辨识方法
CN114416573A (zh) 一种应用程序的缺陷分析方法、装置、设备及介质
Li et al. Logspy: System log anomaly detection for distributed systems
US11227288B1 (en) Systems and methods for integration of disparate data feeds for unified data monitoring
CN114510980A (zh) 模型特征获取方法及装置、电子设备、存储介质
CN114282598A (zh) 多源异构电网数据融合方法、装置、设备及计算机介质
CN113779248A (zh) 数据分类模型训练方法、数据处理方法及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22965615

Country of ref document: EP

Kind code of ref document: A1