With the continuous development and construction of the power system, the source, scale and type of power data continue to expand, control basic data, power grid operation data, primary and secondary equipment monitoring data/status data/working condition data, new energy resource data, natural environment data, etc. The change data and other data together constitute the big data of power regulation. However, these data currently exist independently in the above-mentioned various systems, which are difficult to maintain and share, and lack centralized management and comprehensive analysis of data.
Therefore, in order to further improve the level of power dispatching production management, it is necessary to integrate the data resources of various disciplines and application systems to realize the collaborative sharing of data resources; Development and excavation, and on this basis, research on new power dispatching applications.
1 Power regulation cloud platform based on regulation big data
1.1 Overall Platform Architecture
As shown in Figure 1, referring to the typical structure of cloud computing, the layered design of the power control cloud platform is divided into cloud infrastructure layer (IaaP), cloud middle management layer (IaaP) and cloud application service layer (PaaP) from bottom to top. ).
Application of cloud computing technology in power dispatching
(1) Cloud infrastructure layer: By configuring cluster nodes, disk arrays, switches and other hardware devices, build the underlying architecture of the control cloud platform, and provide data storage, computing and data transmission networks for the control of the middle management layer and application service layer of the cloud platform. Various resources.
(2) Cloud middle management layer: Based on the hardware and resource base of the infrastructure layer, build a real-time data platform, operation data platform, model data platform, etc.
(3) Cloud application service layer: use the data resources and application services of the infrastructure layer and the middle management layer to build application functions based on regulatory requirements.
The power control cloud platform integrates the control data of various business systems, including production control data, management data, etc. Based on the perspective of security protection, the power control cloud platform is deployed in the safe m area.
1.2 Data acquisition and synchronization
1.2.1 Real-time data acquisition
The collection and synchronization of real-time data on the power regulation cloud platform relies on the cloud real-time data platform of the middle management layer of the cloud. Through the online collection at the source end EMP (existing power dispatching control system), the real-time data collection, processing and monitoring of the entire network are realized. storage. The functions of the cloud real-time data platform mainly include data collection and processing, providing data services for upper-layer applications, etc. Mainly as follows:
(1) The data collection function adopts the message forwarding method, obtains real-time data from the source EMP front, and performs message packaging: through the scheduling data network or integrated data network to schedule VPN transmission to the security zone II, and then use the TCP flow method to isolate across the forward direction The device is transmitted to the cloud real-time data platform in the security zone III.
(2) The data processing function provides functions such as data quality processing, analog quantity processing, state quantity processing, and calculation services, and has distributed, sub-regulation mechanism, and sub-regional processing functions.
(3) Measurement model management mainly provides changing measurement models for data collection, collects and transmits changing data in real time, and realizes maintenance-free real-time data cloud platform through measurement model management.
1.2.2 Model data synchronization
Model data synchronization realizes source-side synchronization, splicing, storage, on-demand release and sharing of physical models of the entire network through the model data platform of the cloud middle management layer. According to “Structural Design of General Data Objects for Power Dispatching”, the model data mainly includes public data models, power primary equipment models, automation equipment models, and protection equipment models . The model data platform synchronously summarizes and manages the equipment models and topological relationships of plants and stations of various voltage levels (500kV, 220kV, 110kV, 35kV) in the entire network, as follows:
The model synchronization service is deployed on the data synchronization server of the power dispatching control system in each region of the whole network, and the synchronization event of the model is received.
(1) Each region sends model update information through the bus, sends the exported full and incremental model files to the model synchronization node through the data transmission platform, and releases model update messages at the same time.
(2) The cloud platform returns the model synchronization result and verification report to the region. If the verification fails, the region will be notified for rectification. If the verification is successful, the verified model will be published to the splicing server.
(3) The whole network model splicing program of the cloud platform matches the boundary table with the name of the plant and equipment, obtains the corresponding relationship between the plant, equipment and voltage level ID, and replaces the relevant equipment ID and hot link on the boundary. After the splicing is completed, the spliced model is published to the data synchronization server in the security zone III of the cloud platform through wide-area transmission, and the original relationship database is updated, and the new model is exported and pushed to the cloud, and the application is notified that the new model has been exported successfully.
The network-wide graphic data synchronization method is basically the same as the model data synchronization, and its graphic splicing is based on the model splicing process. After splicing is completed, it is stored in a different database from the model data.
1.2.3 Operation data collection
Based on the full model of the power grid provided by the model data platform, the operation data platform extracts, collects, stores and analyzes the power grid operation data of EMs, OMS, TMR and other business systems, mainly including measurement, power, alarm, event, plan, Seven categories of operating data, including forecast and environment, and related documents are used to support the data analysis and mining applications of the cloud platform. details as follows:
(1) Measurement data: extracted from the real-time database of the real-time data platform, if the extraction fails, the historical measurement data will be retrieved from the relational database.
(2) Electricity data: The electric quantity data of containers or equipment such as power grids and power plants are extracted from the OMS dispatching report in the OMS system, the measurement history data in the relational database, and the collected data of the TMR system, and the source-end system and model data are established The conversion relationship of tables and field attributes between the standard structures of common data objects on the platform.
(3) Alarm data: extract historical alarm data directly from the relational database, and initiate a data recall if the extraction fails.
(4) Event data: The monthly power outage plan, power outage event, equipment failure and other event data are extracted from the OMS system database, and the data that needs to be mapped and converted to the regulatory cloud ID is associated and converted.
(5) Planning data: extract the day-ahead and intra-day plans from the power generation planning module in the safety area II of the power dispatching control system, pack the planning data into a CIME file and send it to the small mail server in the safety area III by small mail, and carry out CIME file parsing and ID mapping conversion.
(6) Prediction data: short-term and ultra-short-term forecasts are extracted from the load forecasting module in the safety zone II of the power dispatching control system, and the forecast data is packaged into a CIME file and sent to the small mail server in the security zone III by small mail, and Perform CIME file parsing and ID mapping conversion.
(7) Environmental data: The environmental data is extracted from the database of the external business system, and saved after ID mapping conversion.
1.3 Regulate big data integration and storage
Various types of data collected and synchronized for the above platforms, such as basic data of the control center/substation, grid steady state/dynamic/transient operation data, monitoring data of power grid dispatching/substation equipment, health status data of substation equipment, electrical Equipment operating condition data, new energy (water, wind, light) resource data, thermal power flue gas emission data, natural environment (meteorology, lightning) change data, etc., the power control cloud platform uses a mixture of traditional relational databases and distributed databases The storage architecture enables integration and storage.
1.3.1 Distributed storage
The hardware architecture of distributed data storage is different from the traditional centralized storage architecture. The hardware of distributed storage adopts virtualized server cluster, and distributes and stores data in multiple storage cluster servers. The servers in the cluster provide cross-redundant storage of data blocks to achieve high reliability and availability of data, and high-speed access to data.
1.3.2 Regulate big data integration
The mutual conversion of unstructured data and structured data is used to support the real-time acquisition of big data, mainly using stream data bus, data converter, and Hive database.
(1) The streaming data bus realizes the collection and processing of massive streaming data, and supports convective data collection and processing before being distributed to each data subscriber. The streaming data bus provides technical support for the monitoring and collection of streaming data such as system logs on the cloud platform.
(2) The data converter realizes the mutual conversion of different structure data, such as the conversion between MysOL and relational database. Since regulatory big data integrates data from different business systems, the source data is stored in different systems using different data structures, and a unified data structure is required after collection.
(3) The Hive database supports the query function of big data, and can establish the mapping between structured data tables and unstructured data files, so as to realize the storage of structured data in Ha4oop unstructured storage.
2 Exploration of the application of power control cloud platform
2.1 Regulate big data analysis index system
Using the traditional data mining method and supplementing and extending it, by combing and counting the existing index system, an index analysis system suitable for fusion data is established based on the massive fusion data.
2.1.1 Sorting out and counting existing indicators
Applying the subject domain and theme data processing concepts of the data warehouse, sorting out various indicators related to the operation and management of various business systems (EMs, OMS, PMs, etc.), and combining the actual assessment indicators of each department and each specialty, a summary is formed Comprehensive indicators across various systems and disciplines, and further form an indicator tree.
2.1.2 Establish a new index evaluation system
Analyze the data sources, statistical cycles, and calculation principles of indicators for each operation management indicator, and establish an indicator evaluation system that integrates various disciplines and systems in accordance with various dimensions such as time and voltage levels.
2.2 Whole process control of maintenance
At present, the whole process of power grid maintenance needs to go through multiple business systems, such as intelligent anti-misoperation ticket system, PMs system, OMS system, etc., and business personnel need to jump to each system when performing the whole process operation: and the data between business systems etc. have not been integrated, such as operation tickets, maintenance work orders, dispatching instruction tickets, etc. are not related to each other, and closed-loop management of the entire maintenance process cannot be carried out. Based on the integration of massive data from various business systems on the power control cloud platform, this paper explores the construction of closed-loop control applications for the entire maintenance process.
2.2.1 Data access and management
(1) Data docking with the EMs system: the EMs system accesses the scheduling instruction ticket, analyzes the operation tasks, operating equipment and operating instruction information, etc., establishes the mapping conversion between the EMs system equipment ID and the cloud platform equipment ID, and performs the scheduling instruction ticket storage.
(2) Data docking with the intelligent anti-misoperation ticket system: the intelligent anti-misoperation ticket system is connected to the sequential control operation ticket and the monitoring operation ticket, and analyzes the operation task, the initial and final status of the equipment, and the information of the operating equipment to complete the intelligent operation ticket The system device ID and the cloud platform device ID are mapped and converted, and the sequence control operation ticket and monitoring operation ticket are stored.
(3) Data connection with the OMS system: the OMS system accesses the maintenance application form and the maintenance application process information, completes the mapping conversion between the OMS system equipment ID and the cloud platform equipment ID, establishes the correlation between the maintenance application form and the dispatching order ticket and store related data.
(4) Data docking with the PMs system: the PMs system accesses the switching operation tickets and maintenance work sheets, and the data access adopts the CIME file or service call method to complete the mapping conversion between the PMs system device ID and the cloud platform device ID, and Store switching operation tickets and maintenance work orders.
2.2.2 Whole-process closed-loop control
Based on the connected forms and information related to the whole process of maintenance, the mutual constraint relationship between each form and process is analyzed, and the control model of the whole process of power grid operation and maintenance is established. It can be divided into the following two process control models.
(1) Power grid operation process management and control model: sort out the operation process and process business data of key links in power grid operation, from writing dispatching order tickets to monitoring operation return orders, and establish a power grid operation process management and control model.
(2) Maintenance business process management and control model: sort out the maintenance process and business data of each key link from the start of maintenance application to the end of maintenance work, and establish a maintenance business process management and control model.
On the basis of the above two management and control models, the status of the whole business process of power grid operation and maintenance is analyzed, and the key process is locked.
2.3 Intelligent arrangement and verification of maintenance plan
The arrangement and execution of the maintenance plan of the power grid is extremely important for the safe and stable operation of the power grid. Based on the power control cloud platform, this paper explores the construction and application of the intelligent auxiliary arrangement and verification of the maintenance plan.
2.3.1 Multi-temporal grid model
In order to support the intelligent arrangement of medium and long-term maintenance plans, it is necessary to build a multi-temporal power grid model first. According to the power grid planning, the power grid planning model for the next few months or one year is modeled, and the model is divided into real-time investment model, near-term preparation for operation model, long-term planning model and historical withdrawal model according to time.
The time attributes of the above-mentioned temporal power grid models will undergo rolling changes in real time with the construction of the power grid, equipment commissioning, and parameter updates. Figure 2 shows this change process. The long-term planning model gradually transforms into a short-term preparation for commissioning model , the model that is ready to be put into operation in the near future has become a real-time investment model, and the new planning model has gradually completed the modeling.
Application of cloud computing technology in power dispatching
The preparation of the maintenance plan can be based on the multi-temporal power grid model to obtain the whole network model of the future time section or a customized model of a specific range.
2.3.2 Arrangement and analysis of maintenance plan
(1) Auxiliary layout of maintenance plan. Based on the establishment of the multi-temporal power grid model, through the analysis of the annual maintenance plan, the monthly maintenance plan and the day-ahead maintenance plan can be automatically generated. On the one hand, it can save the time of maintenance plan planning, and on the other hand, it can avoid maintenance plan adjustments and temporary power outages of equipment. posed risks.
(2) Check the maintenance plan. According to the dispatched maintenance plan check rules, the check strategy of the maintenance plan is formulated, such as construction period constraints, non-simultaneous stop constraints, repeated power outage constraints, and power outage constraints during the power supply period. After the maintenance plan is compiled, it can be checked according to the above rules.
(3) Safety calculation check. As shown in Figure 3, the design idea of safety calculation and verification is as follows: After data integration between the multi-temporal power grid model and the day-ahead/medium- and long-term maintenance plan, the safety verification calculation is performed. The safety check calculation mainly includes power flow calculation and static safety analysis. After the calculation, the safety check calculation results are returned to the maintenance plan for auxiliary arrangement.
Application of cloud computing technology in power dispatching
1) Base state power flow calculation: According to the given grid structure, parameters, and operating conditions of generators, loads and other components, determine the future power grid base state AC power flow distribution, including generator active/reactive output, line power, transformer load, and bus voltage wait.
2) Static safety analysis: judge whether the system will be overloaded or the voltage will exceed the limit after the expected accident occurs. The safety check should have N-1 fault analysis function, and perform N-1 disconnection scanning on all main equipment of the power grid (including lines, main transformers, busbars, and units) including fault sets, and judge whether the system meets the short-term overload capacity after a fault . The scanning of lines, transformers, generators, and DC lines is carried out according to the disconnection of the equipment; switch N-1 can be selected to scan switch maintenance, 3/2 wiring out of strings, disconnection of the bus tie (section) switch, etc., which will cause special problems. In the case of topology changes, bus N-1 only scans the case of single-bus operation for bus maintenance.
With the continuous development of the power system, the diversity and complexity of control data Continuously improving, this paper puts forward the construction idea of the power regulation cloud platform based on the fusion of big data and the storage and integration scheme of massive fusion regulation big data based on cloud computing technology, and explores the construction and application functions of the power regulation cloud platform.