Background
The method has the advantages that the construction of 'two-coverage' is completed, the deepened application of data is realized, a large amount of electric energy data is generated, a new round of marketing and transformation is more targeted, the outstanding problems of low voltage, neck blockage, power failure and the like in poor areas are effectively solved, and the requirements of the masses from 'power on' to 'power on' are met. Meanwhile, the intelligent service level of the power grid is further improved, and a user can enjoy better power supply service. Under the background of the transformation from the current southern power grid to the digitization, the data value after the 'double coverage' is urgently dug, so that the data value can serve for the development of marketing business work, and a solid data base is provided for the intellectualization of inspection business.
In contrast, the chinese patent application with application number CN201911196973.6 proposes a power marketing inspection management method, which includes: acquiring historical marketing service data of the power marketing system, and carrying out standardized processing on the historical marketing service data; analyzing the historical marketing service data after the standardization processing, and extracting abnormal characteristic data; processing abnormal characteristic data based on machine learning and deep learning, establishing an abnormal rechecking positioning algorithm model, completing self-learning of the abnormal rechecking positioning algorithm model, and obtaining an inspection knowledge map; processing the marketing inspection rule, the problem type corresponding to the marketing inspection rule and the data source corresponding to the problem type based on the inspection knowledge graph, positioning abnormal nodes of current marketing service data, obtaining a marketing inspection result, and establishing the inspection knowledge graph for inspection, so that the traceability is realized, and the accurate inspection of the marketing full service, the full data, the full specialty and the full risk is realized.
The Chinese patent application CN110378808A discloses a power marketing inspection method based on gene recombination and feature clustering, which comprises the following steps: step 100, data cleaning, wherein after provincial inspection is carried out on imported work order data, two results are generated after the first data cleaning, namely effective exception and invalid exception of the work order data; step 200, the valid and abnormal data are issued to a city level unit for data processing, and the invalid and abnormal data are directly processed; and step 300, intelligently processing the data processing of the city level units by an intelligent inspection method. The invention can effectively improve the accuracy of on-line inspection and greatly improve the efficiency of on-line inspection operation.
However, the processing mode of the existing marketing system for inspecting the business is strong in subjectivity and insufficient in intellectualization, which not only increases the workload of business personnel, but also cannot meet the requirement of lean management of the inspection.
The current situation of the inspection module of the current marketing management system is specifically analyzed as follows:
1. the rules are not refined enough, and each problem point cannot be positioned in detail;
2. the method is realized based on a relational database, has large performance loss and cannot perform effective parallel processing;
3. the sampling time is long, and the efficiency is influenced;
4. the problems screened out by the rules are more, and the monthly checking traffic is limited and cannot be effectively tracked;
5. the inspection problem and experience are not managed in a knowledge mode, and the problem of the same kind cannot be processed quickly by effectively utilizing knowledge when being processed by different people;
6. the risk monitoring and controlling capability needs to be improved.
Disclosure of Invention
In order to solve the technical problems, the invention provides a risk inspection monitoring platform based on a cloud platform under full-electric-network multi-source data, which comprises an infrastructure layer, a basic software layer, a service supporting layer and an application layer; the infrastructure layer comprises an intelligent electric meter acquisition end and a local edge calculation terminal; the basic software layer comprises database software and a Web middle layer, and provides big data storage service for the risk inspection monitoring platform through the database software and big data basic service for the risk inspection monitoring platform through the Web middle layer; the service support layer provides a big data service suite, and the big data service suite comprises a data set suite, a data governance suite, a query statistics suite, a numerical analysis suite and a data mining suite; and the risk inspection monitoring platform realizes visual intelligent inspection service through the application layer.
The technical scheme of the invention is divided into a data acquisition layer, a data storage layer, a data calculation layer, a data analysis layer and a platform service layer from the flow of data acquisition, storage, communication and use. The platform is integrally deployed on the cloud platform, and the computing and storage services provided by the cloud platform are used, so that the platform naturally has elastic expansion and contraction capability. The platform is a data convergence center, a unique source of unified data and data service, is a development support platform for large data application of each company, and is a carrier for metadata management, data quality management and data standard management.
Various basic technical components required by the project for processing mass data are integrated according to the big data basic platform, the whole life cycle of data processing from data integration, data management, data storage, data query, data statistical analysis and data mining is completely covered, and the necessary analysis technologies of the project such as memory calculation and anomaly analysis are covered.
Specifically, the invention provides a risk inspection monitoring platform based on a cloud platform under full-electric-network multi-source data, and the risk inspection monitoring platform comprises an infrastructure layer, a basic software layer, a service support layer and an application layer.
The infrastructure layer comprises M intelligent electric meter acquisition ends distributed at different positions in a power grid target range and N local edge computing terminals, wherein M > N >1, and M and N are positive integers;
the basic facility layer is communicated with the basic software layer through a distributed file system, the basic software layer comprises database software and a Web middle layer, big data storage service is provided for the risk inspection monitoring platform through the database software, big data basic service is provided for the risk inspection monitoring platform through the Web middle layer, and the big data basic service comprises SQL query, distributed memory calculation, streaming calculation, numerical analysis and data mining;
the service supporting layer provides a big data service suite, and the big data service suite comprises a data integration suite, a data governance suite, a query statistics suite, a numerical analysis suite and a data mining suite;
and the risk inspection monitoring platform realizes visual intelligent inspection service through the application layer.
The application layer also provides a visual configuration service component, and the visual configuration service component sets grouping visual parameters;
based on the grouping visualization parameters, grouping the M intelligent electric meter acquisition ends to obtain K groups, wherein each group comprises a plurality of intelligent electric meter acquisition ends; k is less than or equal to N;
and sending the power grid state data collected by the plurality of intelligent electric meter collecting ends included in each group to at least one local edge computing terminal corresponding to the group.
Each local edge computing terminal is provided with a data cache queue;
and adjusting the packet or data transmission mode by judging whether the data buffer queue is blocked and whether the duration time of the blocking exceeds a preset time value.
The risk inspection monitoring platform also comprises a display layer, wherein the display layer provides an operation interface for a user and realizes the functions expected by the user through the interaction of the user.
In the invention, in order to ensure the safe and stable operation of the marketing system and the high-performance data processing of the inspection abnormity analysis work, the project is realized by combining a big data technology and a J2EE technology. The structure and function modules of each layer are as follows:
infrastructure layer: is the hardware infrastructure layer of the project. For the requirement of intelligent inspection on application, an application server is considered to publish Web application; database servers are contemplated to enable storage of application and management data. And a server cluster is considered to be constructed on the big data platform by adopting the server so as to support big data distributed high-performance mass data processing and analysis.
The basic software layer: for the aspect of intelligent inspection application consistent with the technical architecture of the southern power grid marketing system, the method considers database software and Web middleware and is also consistent with the marketing system in the aspect of selection; mainstream big data platform components are adopted in the aspect of intelligent inspection support, and Hadoop cloud ecological components are mainly adopted. The intelligent rule analysis service provided by each component and the services provided by each component support the requirements of checking archive data processing, business data, table data processing and abnormal rule checking.
A service support layer: designing various business services by adopting a service design concept based on an SOA architecture; and packaging the basic big data development technology in the aspect of intelligent inspection support, and designing a big data service suite.
An application layer: and realizing specific service functions based on the service support layer. For the application planning of the project, the J2EE technology, the micro service technology and the big data technology are respectively adopted for development according to the functional characteristics.
In addition, the invention adopts service-oriented architecture (SOA) to carry out application design, development and system integration, and tries to get rid of the constraint of a technical-oriented solution, thereby improving the reusability of software and accelerating the implementation of application software.
Further advantages of the invention will be apparent in the detailed description section in conjunction with the drawings attached hereto.
Detailed Description
The invention is further described with reference to the following drawings and detailed description.
Referring to fig. 1, a main framework diagram of a risk inspection monitoring platform based on a cloud platform under full-power grid multi-source data according to an embodiment of the present invention is shown.
In fig. 1, the risk inspection monitoring platform includes an infrastructure layer, a basic software layer, a service support layer, an application layer, a presentation layer, and a visual configuration service component.
For further explanation, the risk inspection monitoring platform provided by the invention adopts a provincial level centralized deployment mode, the platform relates to inspection rule statistics, the platform is built on a big data analysis platform, an application server and a database server use virtual resources provided by a cloud platform, and service data come from a cloud database server cluster.
In order to ensure safe and stable operation of the marketing system and high-performance data processing of the inspection anomaly analysis work, the embodiment adopts a mode of combining a big data technology and a J2EE technology.
The platform is structurally divided into: the system comprises four layers of an infrastructure layer, a basic software layer, a service supporting layer and an application layer.
More specifically, on the basis of fig. 1, reference is made to fig. 2.
Infrastructure layer: is the hardware infrastructure layer of the project. For the requirement of intelligent inspection on application, an application server is considered to publish Web application; database servers are contemplated to enable storage of application and management data. And a server cluster is considered to be constructed on the big data platform by adopting the server so as to support big data distributed high-performance mass data processing and analysis.
The basic facility layer is communicated with the basic software layer through a distributed file system, the basic software layer comprises database software and a Web middle layer, big data storage service is provided for the risk inspection monitoring platform through the database software, big data basic service is provided for the risk inspection monitoring platform through the Web middle layer, and the big data basic service comprises SQL query, distributed memory calculation, streaming calculation, numerical analysis and data mining;
the basic software layer: for the aspect of intelligent inspection application consistent with the technical architecture of the southern power grid marketing system, the method considers database software and Web middleware and is also consistent with the marketing system in the aspect of selection; mainstream big data platform components are adopted in the aspect of intelligent inspection support, and Hadoop cloud ecological components are mainly adopted. The intelligent rule analysis service provided by each component and the services provided by each component support the requirements of checking archive data processing, business data, table data processing and abnormal rule checking.
A service support layer: designing various business services by adopting a service design concept based on an SOA architecture; and packaging the basic big data development technology in the aspect of intelligent inspection support, and designing a big data service suite.
The service supporting layer provides a big data service suite, and the big data service suite comprises a data integration suite, a data governance suite, a query statistics suite, a numerical analysis suite and a data mining suite;
an application layer: and realizing specific service functions based on the service support layer. For the application planning of the project, the J2EE technology, the micro service technology and the big data technology are respectively adopted for development according to the functional characteristics.
And the risk inspection monitoring platform realizes visual intelligent inspection service through the application layer.
The application layer also provides a visualization configuration service component that sets grouping visualization parameters.
A display layer: and providing an operation interface for the user, and realizing the function expected by the user through the interaction of the user.
In FIG. 2, the distributed file system employs YARN as the big data cluster resource manager.
Referring next to fig. 3, fig. 3 illustrates in particular one particular arrangement of the infrastructure layer according to the present invention.
Generally, the infrastructure layer comprises M smart meter acquisition ends distributed at different positions in a power grid target range and N local edge computing terminals, wherein M > N >1, and M and N are positive integers.
In fig. 3, 4 smart meter collection terminals and 2 local edge computing terminals are shown.
In fig. 3, based on the grouping visualization parameters, 2 groups are obtained after the 4 smart meter collecting terminals are grouped, and each group includes 2 smart meter collecting terminals;
in FIG. 3, four intelligent electric meter collecting ends are marked as No. 1-No. 4 collecting ends from top to bottom, and No. 1 and No. 3 are divided into a group; no. 2 and No. 4 are divided into one group; if 2 local edge computing terminals are marked as a and B respectively, in the state of fig. 3, the local edge computing terminal a receives the power grid state data collected by the No. 1 and No. 3 intelligent electric meter collecting terminals; and the local edge computing terminal B receives the power grid state data collected by the No. 2 and No. 4 intelligent electric meter collecting ends.
Furthermore, each local edge computing terminal is provided with a data buffer queue.
In each embodiment of the invention, data is processed as locally as possible through the combination of the edge computing terminal and the cloud platform, and if the data is blocked, the grouping is preferentially adjusted; and if the adjustment group cannot meet the capacity expansion requirement, uploading the adjustment group to the platform. The data can be processed in real time by the aid of the arrangement, and meanwhile, rapid increase of cloud transmission cost is avoided.
Specifically, if a certain local edge computing terminal is receiving power grid state data sent by a smart meter collection end, a data cache queue is blocked and the duration time of the blocking exceeds a preset time value, part of the power grid state data collected by the smart meter collection ends in a group corresponding to the local edge computing terminal is sent to the cloud platform.
And if the data cache queue is blocked but the duration time of the blocking does not exceed a preset time value when a certain local edge computing terminal receives the power grid state data sent by the intelligent electric meter collecting end, reducing the number of the intelligent electric meter collecting ends contained in the group corresponding to the local edge computing terminal.
In fig. 3, if a data cache queue of an edge computing terminal a is blocked and a duration time of the blocking exceeds a predetermined time value, then the power grid state data parts acquired by the acquisition ends of the intelligent electric meters 1 and 3 need to be subsequently sent to a cloud platform;
if the data cache queue of the edge computing terminal B is blocked but the duration of the blocking does not exceed the predetermined time value, it may be considered that the state data of the power grid collected by the collecting terminal of the smart meter No. 2 or No. 4 is sent to other edge computing terminals, for example, to the edge computing terminal a; or other local edge computing terminal, or initiate more edge computing terminals, etc.
In specific implementation, if a local edge computing terminal receives power grid state data sent by a smart meter acquisition end, and the data cache queue is blocked but the duration of the blocking does not exceed a preset time value, a feedback signal is sent to the visual configuration service assembly, so that the visual configuration service assembly adjusts the grouped visual parameters.
In the above embodiment, the system further provides a data persistence support service;
after the user sets the grouping visualization parameters through the visualization configuration service component, the service support layer receives a processing request submitted by the user, triggers a service logic, and calls the data persistence support service to complete data storage.
More specifically, see fig. 4.
Packet data is acquired by the infrastructure layer,
calling a JPA tool class through the Web middle layer in the basic software layer to perform data paging query;
the service supporting layer intercepts query data through a data access plug-in and stores the data into a big data cluster server.
The risk inspection monitoring platform provides an Apache CXF WebService running environment, and issues Service services to Web Service to remote control terminals at different positions in the target range of the power grid.
With reference to fig. 4, in this embodiment, various basic technical components required for processing mass data of this project are integrated according to the big data base platform, which completely covers the whole data processing life cycle from data integration, data management, data storage, data query, data statistical analysis, and data mining, and covers necessary analysis technologies of this project, such as memory calculation and anomaly analysis:
wherein:
distributed file system: in order to meet the storage of mass data and the data security requirements, a distributed file system is adopted. HDFS is a highly fault-tolerant system that ensures data security and integrity, and is suitable for deployment on inexpensive machines. HDFS provides high throughput data access and is well suited for application on large-scale data sets.
Cluster resource management: the YARN is used as a big data cluster resource manager, is a universal resource management system, carries out uniform resource management and scheduling on resources such as memory, storage, CPU and the like of related equipment distributed in a cluster, and brings great benefits to the cluster in the aspects of utilization rate, uniform resource management, data sharing and the like.
Big data integration: a plurality of data integration components are provided, and integration of data from different data sources such as relational databases (such as ORACLE |, MYSQL databases), text files (such as meter code txt files, CSV files and the like), instant messages and the like to a large data platform can be realized. The big data platform component mainly adopts a Hadoop cloud ecological component. The data integrity and analysis service provided by each component and the service provided by each component support the requirements of intelligent inspection on archive data processing, business data analysis, table code data processing and rule checking.
1. And the Spark On YARN scheme is adopted to realize the intelligent checking of the base file data and the service data of the users in the whole province in the memory of the cluster according to the checking rule.
2. And (3) large data storage: data warehouse storage, columnar storage, memory-based storage and the like based on the HDFS distributed file system are provided. And the storage requirements of the project mass data are met through various storage modes.
Big data base service: the system comprises a plurality of components for carrying out query statistics, data analysis mining and treatment on data.
More specifically, mapping (ORM) of Java objects and relational databases is realized by means of JPA (JAVA persistent API) technology, and differences of the databases are shielded; the transmission parameter based on the entity type is more beneficial to determining the service interface of the service layer; for JPA-not-good operations, such as association query, high-performance database operation, general entity model definition, and the like, the platform reserves a set of custom data access layers. The database access is shielded through the persistent data service layer, and related operations such as SQL statements, JDBC and the like do not exist on the data service layer.
JPA is a default persistence mechanism of a development platform, and a Spring + JPA + Hibernate implementation mode is specifically adopted. The Entity and the naming query of the JPA are declared with Annotation. The platform provides an Eclipse plug-in, realizes the ORM of the database table to the entity, and organizes the database table to a specific Package for unified management.
The user-defined persistence layer is a data access plug-in provided by the development platform and can be selectively installed according to needs. The plug-in can make up the performance defects of JPA in big data correlation query and batch data processing, is compatible with the processing mode of the original platform, and the custom persistence layer is also the basis of the dynamic entity model.
The service layer receives a processing request submitted by a user, triggers business logic, and calls the service of the persistent layer to complete data storage, and the process is not subject to transaction control. Typically, the service layer is the boundary of a transaction.
The platform provides two transaction control modes: JTA global transactions using containers and declarative transaction control provided by the platform.
The platform provides an Apache CXF WebService running environment. The Service is published as a Web Service, and the following approaches can be provided:
1. adding relevant Annotation in a Service interface by using JAX-WS (Java API For XML Web Service) specifications to change the Annotation into SEI, and then issuing the Service as Web Service by using a platform.
2. And configuring the general Service into Web Service in a registration management module by using a Service registration mechanism to realize configurable release.
The platform supports a variety of mainstream browsers including IE (6.0 and above), FireFox and Google's Chrome. The platform supports browser compatibility by:
1. the basic component and the presentation component provided by the platform realize the compatibility of the browser.
2. When designing functions and pages, the display components (such as labels) provided by the platform should be used as much as possible
3. A Jquery library and a public js function of a cross-browser provided by a platform;
4. for functions with serious difference between browsers, such as a universal modal window, the function is provided by a platform to be realized instead, and the difference between browsers is shielded.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.