CN114493097A

CN114493097A - Power dispatching system operation method based on master-slave mode

Info

Publication number: CN114493097A
Application number: CN202111560303.5A
Authority: CN
Inventors: 邓彬; 郝蛟; 吴睿; 武婕; 王冬; 张思源; 李雨森; 于洋洋
Original assignee: Shenzhen Power Supply Bureau Co Ltd; Nari Technology Co Ltd
Current assignee: Shenzhen Power Supply Bureau Co Ltd; Nari Technology Co Ltd
Priority date: 2021-12-20
Filing date: 2021-12-20
Publication date: 2022-05-13

Abstract

The invention discloses an operation method of a power dispatching system based on a master-slave mode, and aims to solve the technical problem that the cooperative guarantee capability of a standby power dispatching system in the prior art cannot meet the requirements of emergency response and quick transfer. It includes: monitoring and flexibly switching the working mode processes of the 2 power dispatching subsystems; the method comprises the following steps of utilizing a controllable application node to carry out operation control on a power grid system, and collecting data returned by the power grid system; synchronizing model data in the system through master-slave consistency submission and master-slave asynchronous replication operations; synchronizing real-time data of each service in the system by a real-time library synchronization technology; and synchronizing the manual operation data in the system by a message agent forwarding technology. The method and the system can rapidly and accurately manage the data of the power dispatching system in the master-slave mode, improve the cooperative guarantee capability of the standby power dispatching system, and improve the easy maintenance degree of the system.

Description

Power dispatching system operation method based on master-slave mode

Technical Field

The invention relates to a master-slave mode-based power dispatching system operation method, and belongs to the technical field of electrical automation.

Background

The power grid is one of the national important infrastructures, and directly serves national economy and people's life. The safe and stable operation of the power grid is an important guarantee for economic development and social stability. The power dispatching mechanism is a control command center for dispatching and running of a power system, bears important tasks of organizing power grid running, commanding accident handling and recovery, and has very important significance for ensuring safe and stable running of the power grid.

In order to meet the requirement of rapid development of a power grid, prevent the loss of functions of the conventional dispatching center caused by emergencies such as major accidents, natural disasters and the like, ensure the continuity of power dispatching and commanding, improve the emergency capacity of dealing with the emergencies and the reliability of power grid dispatching, ensure the safe and stable operation of the power grid, and urgently need to explore a new dispatching system operation mode.

In order to improve the performance of the power dispatching system, a dispatching system in a master-slave mode is provided, and a traditional power master station master-slave and standby dispatching systems adopt an 'different-address self-standby' operation mode, namely, hardware devices such as a server, an exchanger and a database of the standby dispatching system and master dispatching system devices adopt different addresses. In the operation mode, information such as a basic platform of the standby scheduling system, an SCADA (data acquisition and monitoring control system), a commercial library, a real-time library, a graphic model, a listing and the like under an FES (front-end system) are all synchronized by the main scheduling system, but plant station telecontrol data, alarm information and four remote debugging are all independently acquired by the standby scheduling system; PAS, AVC and other modules of the standby scheduling system also adopt independent operation modes and need to be maintained independently; and the standby dispatching system does not have a planned value issuing function when reaching the unplanned value channel of the power plant, and does not have an interactive channel with provincial and network level dispatching systems.

The construction of the current standby dispatching system mainly has the following problems: 1) the standby scheduling system has a defect in a technical support system architecture; 2) the standby scheduling system technology supports system insufficiency; 3) the normalized application and the operation and maintenance mechanism of the standby scheduling system need to be improved; 4) and the cooperative guarantee capability cannot meet the requirement of rapid transfer.

Disclosure of Invention

In order to solve the problem that the cooperative guarantee capability of the conventional standby power dispatching system cannot meet the requirements of emergency response and quick transfer, the invention provides a master-slave mode-based power dispatching system operation method.

In order to solve the technical problems, the invention adopts the following technical means:

the invention provides an operation method of a power dispatching system based on a master-slave mode, wherein the power dispatching system comprises 2 power dispatching subsystems which are mutually independent and have consistent functional structures, and the operation method comprises the following steps:

acquiring working modes of 2 power dispatching subsystems and monitoring the working modes, wherein the working modes comprise a master mode and a slave mode;

performing control authority management on application nodes in the power dispatching subsystem according to a master-slave mode;

the method comprises the following steps of utilizing a controllable application node to carry out operation control on a power grid system, and collecting data returned by the power grid system;

synchronizing model data in the 2 power dispatching subsystems through master-slave consistency submission and master-slave asynchronous replication operations;

synchronizing real-time data of each service in the 2 power dispatching subsystems by a real-time library synchronization technology;

synchronizing manual operation data in the 2 power dispatching subsystems through a message agent forwarding technology;

synchronizing file data in the 2 power dispatching subsystems through a remote synchronization service;

and switching the working modes of the 2 power dispatching subsystems according to the mode switching instruction.

Further, the method for monitoring the working mode comprises the following steps:

the method comprises the steps that working mode information of each power dispatching subsystem is obtained through a master-slave mode manager deployed on the power dispatching subsystems, wherein the working mode information comprises a system number, a working mode of a system and refreshing time;

each power dispatching subsystem reads configuration information through a master-slave mode manager of the power dispatching subsystem to obtain an address of an opposite-end system proxy server;

each power dispatching subsystem sends own working mode information to the address of the proxy server of the opposite-end system, receives the working mode information of the opposite-end system and collects the working mode information of the 2 power dispatching subsystems.

Further, the method for performing control authority management on the application nodes in the power dispatching subsystem according to the master-slave mode comprises the following steps:

judging whether the subsystem to which the application node belongs is a power dispatching subsystem in the master mode, if so, judging that the subsystem is controllable, otherwise, judging that the subsystem is not controllable;

under the condition that the subsystem is controllable, judging whether the application node is an on-duty node or not, if so, judging that the application node is controllable, otherwise, judging that the application node is uncontrollable;

acquiring control states of all application nodes, and performing global registration of the node control states;

and judging whether the controllable application node is unique, if only one controllable application node exists, configuring a control authority for issuing a control instruction for the application node, and otherwise, changing the control state of the application node into uncontrollable.

Further, the operation of master-slave consistency commit is as follows:

calling a model _ model of the master mode through a service bus, and sending a model operation request to the power dispatching subsystem in the master mode;

the power dispatching subsystem in the main mode executes authority judgment, maximum id generation, model inverse operation statement generation, transaction submission and secondary equipment table triggering information acquisition operation according to the model operation request, and generates a main mode execution result;

if the execution result of the master mode is successful in entering, submitting the trigger information of the secondary equipment table acquired in the master mode in a transaction, and sending a model operation request to the power scheduling subsystem in the slave mode;

if the execution result of the main mode is failure of warehousing, returning an error to a client of the power scheduling subsystem in the main mode;

the power dispatching subsystem in the slave mode executes authority judgment, maximum id generation and transaction submission operation according to the model operation request and generates a slave mode execution result;

and if the slave mode execution result is successful in warehousing, removing the model inverse operation statement generated by the master mode, if the slave mode execution result is failed in warehousing, executing the model inverse operation statement generated by the master mode, and restoring the equipment record in the power scheduling subsystem in the master mode to the state before the model operation request is executed.

Further, the master-slave asynchronous replication operates as follows:

receiving a model operation request by using the power dispatching subsystem in the master mode, and calling a data dispatching service to modify the model and store model data;

generating an archive file according to the model data stored in the master mode and the master-slave mode synchronization target;

calling a data loading process according to the archived file, synchronizing the model data stored in the master mode to the power scheduling subsystem in the slave mode, and returning a slave mode data storage result;

and judging whether to delete the archived file according to the storage result of the slave mode data.

Further, the real-time library synchronization technology comprises timing synchronization, phase data synchronization and snapshot data synchronization, wherein the timing synchronization respectively reads data in databases of 2 power scheduling subsystems according to preset synchronization intervals, and synchronizes the read data to the opposite power scheduling subsystem; the phase data synchronization is triggered by an interface to synchronize the running data in one power dispatching subsystem to the other power dispatching subsystem; and the snapshot data synchronously stores the real-time data in any power dispatching subsystem in a snapshot form and sends the data snapshot to the power dispatching subsystem on the opposite side.

Furthermore, timing synchronization supports 2 power dispatching subsystems to synchronously and relatively synchronize data, and does not support breakpoint continuous transmission; the phase data synchronization is one-way synchronization, and breakpoint continuous transmission is supported.

Further, the method for synchronizing the manual operation data comprises the following steps:

acquiring manual operation information through a power dispatching subsystem, and sending the manual operation information to a local application server and a message agent;

performing application operation by using a local application server according to the manual operation message, and storing local manual operation data;

forwarding the manual operation message to a message agent of the opposite side power dispatching subsystem through a point-to-point service of the message agent;

the manual operation message is forwarded to the application server on the opposite side through a message agent of the power dispatching subsystem on the opposite side;

and performing application operation by the opposite-side application server according to the manual operation message, and storing the manual operation data of the opposite side.

Further, the file data synchronization method comprises the following steps:

acquiring a file directory needing to be synchronized through a file synchronization service of the power dispatching subsystem, and sending the file directory to a remote synchronization service;

changing a synchronous operation log according to the file directory by using a remote synchronous service;

and connecting 2 power dispatching subsystems through a remote agent according to the synchronous operation log to perform file synchronous operation.

Further, the file modification method comprises the following steps: and according to the request of the client, modifying the file in the power dispatching subsystem in the master mode, and storing the modified file by using the file management service.

The following advantages can be obtained by adopting the technical means:

the invention provides a master-slave mode-based power dispatching system operation method, which is used for perfecting the construction of the existing standby dispatching system, provides 2 power dispatching subsystems with the same structural framework and solves the problems of insufficient architecture and incomplete function of the existing standby dispatching system. The method can monitor the working mode of the power dispatching subsystem in real time, provide flexible mode switching, realize seamless switching when the main mode system fails, avoid accident expansion and improve the reliability of power dispatching. In the system operation process, the method provides different data synchronization operations aiming at the synchronization requirements of different data in the system, can provide real-time data synchronization and delay synchronization, ensures the consistency of the data in a master-slave mode system, has real-time property and reliability of the data, improves the cooperative guarantee capability of the power dispatching subsystem in the slave mode, and can meet the requirement of quick transfer.

Part of the operation requirements in the method of the invention are maintained and operated in the master mode, and then the slave mode is synchronously shared, thus effectively meeting the maintenance workload of the whole system and improving the maintainability and easy maintainability of the system. The invention combines the successful construction experience of domestic and foreign dispatching mechanisms and other industrial emergency systems, constructs a safe, reliable, efficient and flexible power grid standby dispatching system, realizes the reliable, efficient and flexible standby of dispatching places, technical support systems and dispatching personnel, meets the omnibearing emergency requirements, and ensures the uninterrupted and reliable operation of power dispatching services.

Drawings

FIG. 1 is a block diagram of a power dispatching system according to an embodiment of the present invention;

FIG. 2 is a schematic diagram illustrating system operation mode monitoring according to an embodiment of the present invention;

FIG. 3 is a schematic diagram illustrating system operation mode switching according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of master-slave coherency commit in an embodiment of the present invention;

FIG. 5 is a schematic diagram of a city model according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a remote mode in an embodiment of the present invention;

FIG. 7 is a flow chart of master-slave consistency commit in an embodiment of the present invention;

FIG. 8 is a flow chart of master-slave asynchronous replication in an embodiment of the present invention;

FIG. 9 is a diagram illustrating real-time data synchronization according to an embodiment of the present invention;

FIG. 10 is an architecture diagram illustrating synchronization of manual operation data in an embodiment of the present invention;

FIG. 11 is a flow chart of manually operating message broker forwarding in an embodiment of the present invention;

FIG. 12 is a flow chart of data proxy forwarding in an embodiment of the present invention;

FIG. 13 is an architecture diagram of file synchronization in an embodiment of the present invention;

fig. 14 is a flowchart of determining control authority of an application node in the embodiment of the present invention.

Detailed Description

The technical scheme of the invention is further explained by combining the accompanying drawings as follows:

the invention provides a master-slave mode-based power dispatching system operation method, wherein a power dispatching system framework oriented to the method is a double-system framework, and the double-system framework comprises 2 mutually independent power dispatching subsystems which can carry out hot standby and have consistent functional structures, and can be called as an A, B system. The framework of the power dispatching system of the invention is shown in fig. 1, and A, B systems all comprise the conventional application functions of the power dispatching system such as preposition, SCADA, alarm, history, calculation, AGC, AVC, PAS, plan, section, graph and the like. A. The working modes of the system B are divided into a master mode and a slave mode, and when the system A is in the master mode, the system B is in the slave mode, and vice versa. In the embodiment of the invention, only the power dispatching subsystem with the main working mode can carry out actual operation control on the power grid, the power grid is not controlled by default in the slave mode, the control is controlled in the slave mode only after part of control right is manually switched, the switching of the control right can be refined to the plant station level, and different plant stations are allowed to execute control in different modes. The working mode can be automatically or manually switched according to actual requirements.

The 2 power dispatching subsystems in the invention can realize the following functions:

(1) in the master mode, the power dispatching subsystem deploys conventional application to be responsible for data acquisition, data processing, historical storage, monitoring and alarming, real-time calculation, online analysis and operation control on the operation state of the power grid;

(2) in the slave mode, the power dispatching subsystem deploys conventional applications to be responsible for collecting, processing, storing, monitoring, calculating and analyzing the running state data of the power grid, online operation can be performed in the slave mode, the operation information is automatically synchronized to the master mode, and the slave mode can select an individual control right or a control right of an individual plant station to control according to needs;

(3) maintaining the master and slave mode models, defaulting to maintaining the models in the master mode, and immediately or in a delayed mode after the maintenance is finished;

(4) the master mode and the slave mode communicate with the lower substation through independent acquisition channels and acquire the operation data of a substation power grid;

(5) the master mode and the slave mode both provide Remote Graphic Service (RGS) for a monitoring workstation to remotely call graphic monitoring interfaces in different modes according to requirements;

(6) the master mode and the slave mode are provided with independent historical databases for historical storage, and the historical databases comprise a relational database, a time sequence database and the like;

(7) the master-slave mode management module is used for managing and switching the master-slave mode, the control right needs to be switched and managed when the master-slave mode is switched, and only the master mode is allowed to carry out power grid control operation;

(8) the operation (manual setting and listing) information under the master-slave mode is automatically synchronized, and the consistency of the operation information between the master-slave mode is ensured. When the model between the master mode and the slave mode adopts delay synchronization, only the operation in the master mode is allowed during the delay release period, and the operation information of the master mode is automatically synchronized in batch when the delay is due and synchronized, and the slave mode can be compared with the existing operation information to judge the incremental operation to be put in storage.

The method mainly comprises a system working mode management method, a system data synchronization method and a system control authority management method, wherein the system working mode management method is responsible for managing the working modes of 2 power dispatching subsystems, and only one subsystem is ensured to run in a master mode under the normal running condition; the system data synchronization method synchronizes various data information in the 2 power dispatching subsystems, and the system control authority management method manages the application node control authority in the 2 power dispatching subsystems, so that only the subsystems in the master mode can carry out power grid dispatching operation.

The system working mode management method comprises the following steps:

and step A, monitoring the working modes of the 2 power dispatching subsystems.

A master-slave mode manager is arranged in the power dispatching subsystem, and the working mode monitoring function is realized by interacting respective working mode information in real time through 2 subsystems, as shown in fig. 2, the specific operation of the working mode monitoring is as follows:

step A01, obtaining the working mode information of each power dispatching subsystem through a master-slave mode manager deployed on the power dispatching subsystems, wherein the working mode information comprises a system number (A or B), a working mode (master mode or slave mode) of the system, refreshing time and the like.

And step A02, each power dispatching subsystem reads the configuration information through the master-slave mode manager to obtain the address of the proxy server of the opposite-end system.

Step A03, each power dispatching subsystem sends its own working mode information to the address of the proxy server of the opposite terminal system, receives the working mode information of the opposite terminal system from the opposite terminal system, and collects the working mode information of 2 power dispatching subsystems.

The invention can check the refreshing time of the working mode information while monitoring the working mode of the system, if the working mode information is not refreshed after overtime, the invention can carry out alarm notification, and a user can check whether the abnormity of network interruption between systems, opposite end system fault and the like occurs according to the alarm information so as to implement fault treatment measures. When the network between the systems is interrupted and the 2 subsystems of the invention are split into two systems, the system working mode monitoring sends out a system interruption alarm, and the user determines the master-slave working mode of each subsystem. When a certain subsystem is destroyed under the conditions of natural disasters and the like, the system working mode monitoring sends out a system interruption alarm, and a user determines the master-slave working mode of the subsystem according to the destruction degree.

And B, switching the working modes of the 2 power dispatching subsystems according to the mode switching instruction.

In the normal mode switching process, the master-slave mode manager of the power scheduling subsystem in the master mode receives a mode switching instruction from a client or a system server, updates the working mode of the system according to the mode switching instruction, simultaneously sends the updated working mode of the system to the opposite system, and the master-slave mode manager of the opposite system automatically switches the working mode of the master-slave mode manager to a corresponding mode according to the received information, as shown in fig. 3, the system a is the master mode, after the system a obtains the mode switching instruction, the working mode of the system a is firstly switched from the master mode to the slave mode, then the updated slave mode of the system a is sent to the system B, and finally the working mode of the system B is switched from the slave mode to the master mode.

In the embodiment of the invention, the working modes of the system comprise an enforced master mode and an enforced slave mode besides the master mode and the slave mode. The mandatory master mode means that the subsystem is manually switched to the master mode and the other subsystem is automatically switched to the slave mode. Similarly, forcing the slave mode means that the subsystem is manually switched to the slave mode and the other subsystem is automatically switched to the master mode.

In the process of forcibly switching the modes, when a master-slave mode manager of a certain subsystem receives a switching instruction to switch the subsystem into a master mode, the master-slave mode manager of the subsystem firstly switches the working mode of the system into the forcibly master mode and then sends the working mode of the subsystem to a counter-end system; the opposite end system finds that the working mode of the original system is a mandatory master mode, automatically switches the working mode of the opposite end system into a slave mode, and sends the working mode to the original system; the original system finds that the working mode of the opposite-end system is the slave mode, switches the working mode from the mandatory master mode to the master mode, and completes the mode switching. Similarly, the subsystem may perform a forced switch from mode in a similar manner.

In the embodiment of the invention, according to different deployment positions of the 2 subsystems, the power scheduling system can be a living system in the same city or a remote living system, the 2 power scheduling subsystems in the living system in the same city are deployed in the same city, and the 2 power scheduling subsystems in the remote living system are respectively deployed in different cities. Since 2 subsystems are running simultaneously, they are hot spares of each other. The system data synchronization of the same city double-activity system or the different place double-activity system is different.

The system data synchronization method comprises the following steps:

and step C, synchronizing the model data in the 2 power dispatching subsystems through master-slave consistency submission and master-slave asynchronous replication operations.

The model data synchronization is divided into master-slave consistency submission and master-slave asynchronous replication; the master-slave consistency submission is that data are written into a subsystem database in a master-slave mode by a model service program, if the data are submitted successfully in the master-slave mode, the model operation is formally submitted successfully, otherwise, the master-slave mode is returned to the pre-operation state through the model rollback, and the real-time consistency of the master-slave model data is ensured through a strong consistency technology; the master-slave asynchronous replication mode asynchronously synchronizes the model data in the master mode to the commercial library corresponding to the slave mode through a commercial library service program provided by a commercial library component or a platform, so that the final consistency of the model data of the system in the master-slave mode is ensured.

In the embodiment of the present invention, the working principle of master-slave consistency submission is shown in fig. 4, which specifically includes:

step C1-1, calling a model service interface of the system through the client to generate a request message for requesting model operation; the client is a power dispatching system client deployed on a workstation, the client is used for providing a main interface for operation for dispatching personnel, and the processing and operation of the graph modulus on the server, such as the operation of modifying the model, are completed through the operation of the front-end interface of the client.

And step C1-2, calling the model _ model of the master model through the service bus by the client model operation request, wherein the process is mainly a network transmission process, and sending the client model operation request to the power dispatching subsystem in the master model.

And step C1-3, the power dispatching subsystem in the main mode executes operations such as authority judgment, maximum id generation, model inverse operation statement generation, transaction submission, secondary equipment table trigger information acquisition (namely statements generated by the secondary equipment table trigger information) and the like according to the client model operation request, completes the core stage of service processing of the main mode model and generates a main mode execution result.

And step C1-4, the power dispatching subsystem in the master mode returns the execution result of the master mode to the area system model _ model in the form of a message. The regional system is a system which is independent of the master-slave mode and is used for coordinating the operation conditions of all participants, and is respectively connected with the power dispatching subsystem in the master-slave mode through a network. Step C12 and step C14 are the request and return phases in a service bus call.

Step C1-5, returning message content by analyzing the master mode through the regional system, and returning an error to the client if the master mode fails to be put in storage; and if the primary mode is successfully put in the database, locally submitting the secondary device table triggering information acquired by the primary mode in a transaction.

And step C1-6, the regional system model _ model forwards the master mode return message content and the model operation request to the slave mode model _ model, the power scheduling subsystem in the slave mode performs operations such as permission judgment, maximum id generation, transaction submission and the like according to the client model operation request, completes the core phase of slave mode model service processing, and generates a slave mode execution result.

And step C1-7, the slave mode model _ model sends the slave mode execution result to the master mode model _ model in the form of a message through the service bus.

Step C1-8, the master mode analyzes the message returned from the slave mode, if the slave mode fails to be put in storage, the master mode executes the pre-written model inverse operation statement, and the equipment record in the master mode is restored to the state before modification; and if the slave mode is successfully put in the database, the master mode clears the pre-written model inverse operation statement.

Step C1-9, the processing result (execution or clear model inverse operation statement) of the model inverse operation statement by the main mode returns to the region system model _ model.

And C1-10, analyzing the processing condition of the main mode on the inverse operation statement by the regional system, and returning to the calling client.

The submission statement in the initial data actual submission phase of the master schema model service contains the following information: 1. an inverse operation backspacing statement of the operated record id; 2. recording a mark record of the successful submitted id 1 in the model inverse operation information table; 3. the main key id of the current session in operation is used for generating backspacing statement information of a triggered table in a trigger; 4. record the actual operation commit _ sql statement for id.

The four types of statements are submitted in one transaction begin … end, so that the successful submission of the comment _ sql is guaranteed, and the mark record of success _ flag ═ 1, the operated record and the rollback statement affecting the triggered table in the public _ model _ undo table successfully warehoused by the model center are all submitted successfully.

In the same city dual-activity system, the A, B system can share one commercial library, the commercial library comprises databases with different roles such as a main library, a standby library and the like, the specific structure is shown in fig. 5, and the single-point failure problem of the databases is solved through multiple roles in one commercial library cluster. The platform commercial library service module is responsible for the management functions of the main database and the standby database, the main system and the slave system submit query data to the main library according to data requirements, and the commercial library synchronization function completes the synchronization of the standby database.

In the different-place dual-active system, the A, B system corresponds to a commercial library, and the commercial libraries of the A, B system and the commercial libraries of the two systems are maintained respectively, as shown in fig. 6, model operation can adopt master-slave consistency submission or master-slave asynchronous replication to realize synchronization, and internal data of the system, such as historical data, and the like are stored independently at two sides. The master-slave consistency submission is used for scenes with higher data consistency requirements, such as model data maintenance; master-slave asynchronous replication is used for sampling, alarm data and other scenes.

The master-slave consistency submission is realized based on a three-phase submission protocol of the data service, the three-phase submission depends on the role of a coordinator, the coordinator is a set of scheduling services deployed in two subsystems, and A, B is a commercial library of the system and serves as a participant cluster. The coordinators are uniformly scheduled to send instructions to all the participants, the participants can feed back the operation condition to the coordinators, and the coordinators determine the next step of instructions according to the feedback condition. The specific operation flow of master-slave consistency submission is shown in fig. 7:

(1) and an inquiry stage: and sending a submission inquiry request to all the commercial library services through the dispatching service for confirming whether the commercial library of each power dispatching subsystem can normally respond or not, wherein the commercial library returns a positive response if the submission is available, and otherwise, returns a negative response.

(2) And a pre-submission stage: the scheduling service determines whether a pre-commit operation can be performed according to a feedback condition of an inquiry stage, and specifically includes the following two conditions:

(201) if the dispatch service receives a positive response from both A, B systems: and performing pre-submission of a transaction, wherein the transaction refers to a model modification request for the commercial library, which is sent by a client, and the scheduling service sends a pre-submission request to all commercial library services and enters a preparation phase. After receiving the pre-commit request, the commercial library service executes the transaction operation and records the undo and redo information into a transaction log. If the business library service of one subsystem successfully performs the transaction operation, a positive response is returned while beginning to wait for the final instruction, and if the business library service cannot perform the transaction operation, a negative response is returned.

(202) If the dispatch service receives at least one negative response or no response is received after waiting for a timeout: and interrupting the transaction, wherein the scheduling service sends an interruption request to commercial library services of the two subsystems. And executing transaction interruption if the commercial library service does not receive the scheduling service request after receiving the interruption request from the scheduling service or after time-out.

(3) And a formal submission stage: the scheduling service determines whether to continue the formal commit operation of the transaction according to the response of the participant in the pre-commit phase, which may be specifically divided into the following two cases:

(301) if the dispatch service gets a positive response from all subsystems: the scheduling service enters the commit state from the pre-commit state and sends a formal commit request to the commercial library services of all subsystems. And after receiving the formal commit request, the commercial library service executes formal transaction commit, releases all transaction resources after completing the transaction commit, and sends a haveCommitted positive response to the scheduling service.

(302) If the dispatch service receives at least one negative response or no response is received after waiting for a timeout: the transaction is interrupted, the specific operation is the same as the interruption in the pre-commit phase.

(4) And after the dispatching service receives the positive responses submitted by all the subsystems formally, the task is completed.

A. The master-slave asynchronous replication function of the commercial database data of the system B is realized based on the data submission of the platform archive file, and the main process is as shown in FIG. 8:

and step C2-1, receiving the model operation request by the power dispatching subsystem in the master mode, calling data dispatching service by the client, modifying the model and storing the model data, returning the model data immediately after the data is successfully stored, and avoiding subsequent data synchronization when the data is unsuccessfully stored.

And step C2-2, generating a plurality of transaction-level archive files according to the system configuration and the synchronization target of the master-slave mode through the data scheduling service, and synchronizing data to the power scheduling subsystem in the slave mode.

And step C2-3, monitoring the generation of the archived file through the archived file capturing process, and once the generation of the archived file is found, synchronously calling a data loading process of the remote center through high-speed resource network multithreading.

And step C2-4, the data loading process synchronizes the model data saved in the master mode to the power scheduling subsystem in the slave mode according to the content of the archive file, completes the saving of the data, and returns the saving result of the slave mode data to the archive file capturing process.

And step C2-5, the archive file capturing process obtains the slave mode data storage result according to the data loading process of the remote center, and further judges whether to delete the local archive file.

When the power dispatching subsystem in the slave mode normally stores data, deleting the archived file; when the power dispatching subsystem in the slave mode sends an exception, such as: if the network is abnormal, the data storage is abnormal and the data loading process is abnormal, the archive file capturing process does not delete the archive file; and when the power dispatching subsystem in the slave mode saves the internal logic error, moving the archive file to an error directory and recording the error information saved by the slave system for subsequent processing.

And D, synchronizing the real-time data of each service in the 2 power dispatching subsystems by using a real-time library synchronization technology.

Real-time library synchronization provides the ability to synchronize real-time data of various services from the a system to the B system. A. Although the systems B are mirror images of each other, the processing strategies of the services are different, and the understanding of the data synchronization requirements and consistency is different. Taking control type service as an example, AGC and AVC service require A, B service between systems to work completely independently, A, B service between systems does not interact, and consistency depends on program synchronization processing in consistent SCADA and state estimation result. Meanwhile, the plan service considers the characteristics of the processing logic, requires the main system to process, receives synchronization by the standby system, and takes over the service under the condition that the main system is abnormal, so all stage data which is formed by the main system and affects the subsequent processing need to be immediately synchronized to the standby system for storage. Meanwhile, similar to offline analysis services, such as dispatcher load flow calculation, a service instance may only be started in the system A, the processing result and related data of the service instance are synchronized to the system B for storage, and when the system A fails, recovery is performed according to backup data stored in the system B.

Because the demands of the services for real-time data synchronization are different greatly, the invention provides different data synchronization strategies and mechanisms for real-time data, and the real-time data are selected and used by the application autonomously, as shown in fig. 9, specifically including timing synchronization, phase data synchronization and snapshot data synchronization.

1. Timing synchronization

Real-time data timing synchronization mutual synchronization of designated data sets is accomplished between A, B systems at a fixed time, the synchronized data sets being capable of being set at three levels, tables, fields and records. The timing synchronization program reads real-time data from the database of the A, B system according to the data set configuration at the set synchronization time and synchronizes to the contralateral system. The timing synchronization supports A, B the system and the same table are synchronous at the same time, so the system is suitable for the synchronization of the monitoring data of the system, and the timing synchronization does not carry out breakpoint transmission considering the continuity of the timing synchronization.

2. Phase data synchronization

The phase data synchronization provides a data synchronization means triggered by application, the application in the subsystem triggers the synchronization of the specified data set through an interface, and the synchronization of data from one subsystem real-time library to another subsystem real-time library can be completed in real time. Phase data synchronization supports table-level, field-level, and record-level synchronization granularity. Phase data synchronization needs to be synchronized from the running subsystem to the standby subsystem of the application, so that only one-way synchronization is provided, and the same table does not consider two-way synchronization. The phase data synchronization supports the breakpoint continuous transmission function, and the application can select a synchronization strategy with the breakpoint continuous transmission function.

3. Snapshot data synchronization

Snapshot data synchronization provides offline data cross-site data backup capability, and an application can choose to store a snapshot of the running data synchronously to another subsystem. When one subsystem is unavailable, an application can be recovered in another subsystem through offline snapshot, and the method is mainly used for fast migration and recovery of offline analysis services.

The following table compares three real-time data synchronization modes:

and E, synchronizing the manual operation data of the 2 power dispatching subsystems through a message agent forwarding technology.

The manual operation data synchronization technology realizes the automatic synchronization of the user operation information of the same city double-activity system or the different place double-activity system through the message agent forwarding technology among the systems. The same-city dual-active system determines master and slave positions through a load balancing strategy, operation information is sent to a host system SCADA application and is respectively synchronized to a standby system through information agents, and therefore operation data are determined to keep synchronization at two-place nodes. And the backup system and the main system finish the same operation steps after receiving the synchronous message, and write the information into the local real-time database. For the same-city double-activity mode, the judgment of writing library permission needs to be added in the standby system, the consistency of commercial library data is determined, and data repeated writing is avoided.

The types of operation synchronization comprise operations such as setting, blocking, listing, banning and suppression, and the operation synchronization granularity comprises equipment, interval, station and full-graph operations.

In the abnormal state of the intermediate network or the proxy of the two-place system, a message cache mechanism is adopted to cache the operation message to be forwarded to the local, and the local cache message is obtained again for forwarding processing after the system communication is recovered to be normal.

As shown in fig. 11, the power grid models and message structures of the main system (power dispatching subsystem in the master mode) and the standby system (power dispatching subsystem in the slave mode) are completely the same, the two systems are connected through the message broker, the communication between the message brokers can be processed through safety devices such as a firewall or an encryption facility, and the user operation synchronization of the mutual standby system is realized through the message broker forwarding technology between the systems.

And E01, setting a message agent in each of the 2 power dispatching subsystems for receiving the operation messages from the system in real time.

And step E02, when the user operates on the picture of any power dispatching subsystem, the manual operation message is sent to the local application server and is also sent to the local message agent.

And step E03, the message agent forwards the manual operation message to the opposite message agent through the point-to-point service, and the manual operation message may be transmitted by a security device such as a data encryption platform or a network firewall in the middle.

And E04, the opposite side message agent machine forwards the manual operation message to the application server of the opposite side system, so that the complete real-time synchronization of the manual operation data is realized.

If a single-side system failure occurs in step E03, after the single-side system failure is recovered, manual operation data synchronization is implemented by using an intersystem data table synchronization technique, as shown in fig. 12, two sets of systems that are mutually equipped are connected by a data agent, communication between the data agents can be processed by a security device such as a firewall or an encryption facility, and the operation data synchronization is implemented by using the intersystem data table synchronization technique, which is suitable for data recovery after the single-side system failure.

And step E05, configuring the data table to be synchronized through the configuration file, wherein the data table includes a table name and a synchronization mode, and full table synchronization or partial field synchronization can be configured.

And E06, respectively arranging a data proxy server in the main system and the standby system for receiving the data from the system in real time.

And E07, after the single-side system is recovered from the fault, reading the configuration file by the application server host in the system which needs to synchronize to the opposite side, packaging the content of the specified data table as required, and sending the data table to the data proxy server. For example, if the system B fails and data synchronization needs to be performed through the system a after recovery, the application server of the system a reads the configuration file, and packages the specified data table and sends the data table to the data proxy server of the system a.

And step E08, the data proxy server packages and sends the data to the data proxy server of the opposite side system in a point-to-point mode, and the data can be transmitted by a security device such as a data encryption platform or a network firewall in the middle.

And E09, after the data proxy server of the opposite side system receives the packaged data, analyzing the data content and writing the data content into the active/standby machine real-time library of the application server, so as to realize the full data synchronization of the data table of the opposite side system.

And F, synchronizing file data in the 2 power dispatching subsystems through the remote synchronization service.

The file service is arranged in 2 power dispatching subsystems, the file service provides file and directory management functions for a local system, the file service generally adopts one master-one-standby configuration, a host receives a client request, a file is stored in the local system, after the file is modified, modification operation and modification content are sent to a file service standby machine through system internal synchronization service, and the data consistency of the file of the two machines in the system is realized.

Under the same city dual-active mode and the remote disaster recovery mode, the file service cannot directly communicate with a system in the same city or in a remote place, and remote synchronization needs to be realized based on the proxy service. The remote synchronization service records the remote synchronization log by configuring the directory which needs remote synchronization, and the remote synchronization service performs incremental file synchronization through a remote agent according to the file change log, so that the operation is redone in a remote system, and the consistency of file data among A, B systems is realized. The file synchronization architecture is shown in fig. 13.

In order to meet the requirement of simultaneously modifying files in the disaster recovery system of the same city, double activities and different places, the file synchronization service of the invention needs to support a bidirectional synchronization function. In the dual active mode, the file management services of the A, B systems simultaneously serve the applications of the respective systems for storing and managing files of the respective systems. For the file data only used in the system, the synchronization between systems is not needed, only the file data is stored in the local system, and only the file data which needs to be shared between A, B systems is synchronized.

To implement the bi-directional synchronization function, the file management service needs to distinguish between the locally requested and remotely synchronized files. The client requests the data message to increase an identifier, when the local client requests, the identifier is local, and when the remote synchronization is carried out, the identifier is set to be remote, so that the file service does not record a remote synchronization log for the remote synchronization request, and the cyclic synchronization is avoided.

In the A, B system, the same file cannot be modified simultaneously, otherwise the final synchronization result is unknown. In order to solve the problem, the method adopts a master mode for control, when one type of files have the possibility of simultaneous modification conflict, the files can be uniformly modified in a master mode system, and the modified files are synchronized to a slave mode system by the master mode system, so that the conflict is avoided; the file can also be locked in the master mode system, and after locking, the slave mode system fails to request locking, thereby avoiding the conflict of simultaneously modifying the same file.

When files are remotely synchronized, the problem of network failure between systems needs to be considered A, B, so that a synchronization operation log is not lost during the failure period, and the files are continuously synchronized after the failure is removed. The method adopts a persistence mode to store the synchronous operation logs on the file service machine according to the sequence of the requests, once A, B system network fails, the synchronous operation logs are stored according to time, and after the failure is relieved, the file synchronization service synchronizes the files in sequence according to the time of the synchronous operation logs until the synchronization is completed, thereby realizing the final consistency of the file data.

In the embodiment of the invention, the control type application (such as SCADA, AGC, AVC and the like) in the power dispatching system adopts an overall mode of independent operation, bidirectional synchronization and unique control, and realizes a dual-active operation control mode of a master system and a slave system. The application functions are deployed and independently run on the master and slave system sites in a master-slave mode, each application node depends on platform system services and database management of the system, model parameters and real-time measurement data are obtained from the system, and independent analysis and calculation are carried out.

In order to meet the requirements of dual activities, all model maintenance, parameter maintenance, interface operation and the like need to be synchronized in two directions through a platform message mechanism, the consistency of a calculation model and parameters is kept in real time, and the mode switching of a master system and a slave system at any time is met. The application operation of the unified man-machine is firstly sent to the application node of the main system, and then the operation data are respectively synchronized to the rest nodes of the slave system through the message agents, so that the synchronization of the operation data at the nodes of the main system and the slave system is ensured. And after receiving the synchronization message, the master-slave system completes the same operation steps, thereby ensuring the real-time synchronization of the operation and maintenance of the master-slave system.

Step G, in order to ensure the safety of the control system, the method of the present invention needs to perform controllable state judgment and mutual check on each application node of the power scheduling subsystem, and ensure the uniqueness of the control node at the same time, as shown in fig. 14, the specific operations are as follows:

and G01, judging whether the current system is controllable through the application nodes, wherein only the power dispatching subsystem on-duty application host in the master mode is controllable, and the rest is uncontrollable.

And G02, if the current system is controllable, judging whether the application node is the on-duty node, if so, controlling the application node, otherwise, not controlling the application node.

And G03, performing global registration of node control states by each application node, namely counting whether all the application nodes in the system are controllable.

G04, performing control right check on the controllable nodes, judging whether the application node is unique, if only one controllable application node exists, checking that the application node passes the check, and if not, changing the control state of the application node into uncontrollable.

And step H, whether in the same city or in different places, the master system and the slave system have complete data acquisition application, are respectively communicated with all the managed stations, and have the function of issuing control commands. The data acquisition application does not have the requirement of actively sending the control command to the plant station, and only converts the received control commands of other applications in the system into protocol messages to be sent to the plant station. For the safety of system operation, after receiving control commands issued by other applications in the system, the data acquisition application judges whether the subsystem to which the node belongs is a 'power scheduling subsystem in the master mode', if so, the data acquisition application allows the control commands to be issued, otherwise, the data acquisition application does not allow the control commands to be issued, thereby ensuring the safety and uniqueness of the issuing of the control commands.

In the embodiment of the present invention, the numbers a to H are only used to distinguish different operation steps, and the operation order of each step is not limited.

The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims

1. A power dispatching system operation method based on a master-slave mode is characterized in that the power dispatching system comprises 2 power dispatching subsystems which are independent from each other and have consistent functional structures, and the operation method comprises the following steps:

2. The method for operating the power dispatching system based on the master-slave mode as claimed in claim 1, wherein the method for monitoring the working mode comprises:

3. The method for operating the power dispatching system based on the master-slave mode as claimed in claim 1, wherein the method for performing control authority management on the application nodes in the power dispatching subsystem according to the master-slave mode comprises:

4. The method for operating the power dispatching system based on the master-slave mode as claimed in claim 1, wherein the master-slave consistency submission is performed as follows:

respectively calling model _ models of the master mode and the slave mode through a service bus, and sending model operation requests to 2 power dispatching subsystems;

if the execution result of the main mode is successful in warehousing, submitting the secondary equipment table triggering information acquired by the main mode in a transaction, and if the execution result of the main mode is failed in warehousing, returning an error to a client of the power scheduling subsystem in the main mode;

5. The method for operating the power dispatching system based on the master-slave mode as claimed in claim 1, wherein the master-slave asynchronous replication is performed as follows:

receiving a model operation request by using the power dispatching subsystem in the master mode, and calling a data dispatching service to modify the model and store the model data;

calling a data loading process according to the archive file, synchronizing the model data stored in the master mode to the power scheduling subsystem in the slave mode, and returning a slave mode data storage result;

6. The power dispatching system operation method based on the master-slave mode as claimed in claim 1, wherein the real-time library synchronization technology comprises timing synchronization, phase data synchronization and snapshot data synchronization, wherein the timing synchronization reads data in databases of 2 power dispatching subsystems respectively according to a preset synchronization interval and synchronizes the read data to the opposite power dispatching subsystem; the phase data synchronization is triggered by an interface to synchronize the running data in one power dispatching subsystem to the other power dispatching subsystem; and the snapshot data synchronously stores the real-time data in any power dispatching subsystem in a snapshot form and sends the data snapshot to the power dispatching subsystem on the opposite side.

7. The power dispatching system operation method based on the master-slave mode as claimed in claim 6, wherein the timing synchronization supports 2 power dispatching subsystems to simultaneously synchronize data relatively, and does not support breakpoint transmission; the phase data synchronization is one-way synchronization, and breakpoint continuous transmission is supported.

8. The power dispatching system operation method based on the master-slave mode as claimed in claim 1, wherein the manual operation data synchronization method is as follows:

9. The power dispatching system operation method based on the master-slave mode as claimed in claim 1, wherein the file data synchronization method is:

10. The power dispatching system operation method based on the master-slave mode as claimed in claim 9, wherein the file modification method is: and according to the request of the client, modifying the file in the power dispatching subsystem in the master mode, and storing the modified file by using the file management service.