CN109167354A - A kind of power grid forecast failure parallel parsing calculation method based on exchange files - Google Patents
A kind of power grid forecast failure parallel parsing calculation method based on exchange files Download PDFInfo
- Publication number
- CN109167354A CN109167354A CN201811166572.1A CN201811166572A CN109167354A CN 109167354 A CN109167354 A CN 109167354A CN 201811166572 A CN201811166572 A CN 201811166572A CN 109167354 A CN109167354 A CN 109167354A
- Authority
- CN
- China
- Prior art keywords
- resource
- task
- calculating
- computing resource
- failure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/088—Aspects of digital computing
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Power Engineering (AREA)
- Debugging And Monitoring (AREA)
Abstract
The electric network fault parallel parsing calculation method based on exchange files that the present invention relates to a kind of, technical characterstic are: the following steps are included: step 1, establishing task list, computing resource table, establishing system failure collection；Computing resource information, calculating task information are established according to computing resource server is participated in；Step 2 runs total task schedule process in task scheduling server, when every wheel static security analysis function cycle of operation arrives, increase new power grid static security analysis in task list and calculates general assignment, and the specific parallel computation task of each computing resource is resolved into according to the general assignment that the forecast failure of static security analysis calculates, it is written in computing resource table；Step 3, on computing resource server, computing resource finger daemon；Step 4, in task scheduling server, wait next static security analysis calculating cycle to arrive.The present invention can be improved the on-line analysis calculating speed of static security analysis function.
Description
Technical field
The invention belongs to technical field of power systems, it is related to power grid forecast failure parallel parsing calculation method, especially one
Power grid forecast failure parallel parsing calculation method of the kind based on exchange files.
Background technique
Power grid static security analysis is one of the main application software of modern power network operation control centre, and main target is
It needs to preset the failure that may occur in power grid according to personnel, passes through the operation of power networks after each failure occurs for line
Situation carries out analytical calculation, the accurate system operation mode to after being out of order, and intuitively accurately shows various fail results, thus
Quickly judge the harm that various failures generate operation of power networks and in time prompt operator acquisition prevention mistake.Power grid is quiet
The essential core function of state safety analysis is to carry out the various electric network faults of anticipation quickly in line computation, and calculated result provides pre-
Think overload branch under failure, electric voltage exception bus, pressure anomaly node and out-of-limit generator, steam power plant etc., and provides
Its out-of-limit degree provides analytical calculation foundation to ensure that power system mesomeric state is safe and reliable to operation.
With the fast development of Electricity Demand to maintain sustained and rapid growth with power industry, power grid scale constantly expands.
On the extensive of recent domestic large-scale blackout studies have shown that influencing huge large-scale blackout mostly and being by static security about
Beam destruction is gradually evolved: circuit on power system and transformer etc. are transmitted electricity, and element failure is out of service, and initiation power is big
Range transfer, causes adjacent elements to overload, and forms chain over-load trip, eventually leads to large area blackout.Therefore, power train
Forecast failure in system static security analysis calculates the important means for having become the safe operation for guaranteeing electric system.
Common electric network fault analysis calculation method is using tidal current computing method in static security analysis.Load flow calculation is
According to the service condition of the elements such as given electric network composition, parameter and generator, load, determine that electric system each section stable state is transported
The calculating of row state parameter.Usually given service condition has each power supply and the power of load point in system, pivot point voltage, puts down
The voltage of weighing apparatus point and phase angle.Running state parameters to be asked include the voltage magnitude and phase angle of each bus nodes of power grid, and
The power distribution of each branch, power loss of network etc..Accident analysis be exactly after the initial operating condition of given Load flow calculation,
It is out of service that the failures such as the forecast failure influence to service condition, such as route, main transformer, generator are given again, are then carried out again
Load flow calculation, to obtain the power of a bus nodes voltage and branch after forecast failure occurs.
Currently, the monitoring system of dispatching of power netwoks control centre above the provincial level all has static security analysis function.Online fortune
Scan period of the capable static security analysis in control centre is generally 1-15 minutes, and it is 1 minute that majority, which requires the period, therefore is compeled
The calculating speed of online static security analysis will be improved by being essential；Meanwhile with the fast development of China's power grid, control centres at different levels
The power grid scale of monitoring is increasing, and the calculate node number in power grid computation model is online to carry out in advance mostly more than 4000
Think that the number of faults of calculation of fault alreadys exceed 8000, it is so large-scale to calculate to online contingency analysis calculation method
Speed it is also proposed that requirements at the higher level, the existing method of tradition cannot meet the needs.
Summary of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of power grid forecast failure based on exchange files
Parallel parsing calculation method improves the on-line analysis calculating speed of static security analysis function.
The present invention solves its realistic problem and adopts the following technical solutions to achieve:
A kind of electric network fault parallel parsing calculation method based on exchange files, comprising the following steps:
Step 1, before calculating, establish task list, computing resource table, establish system failure collection；According to participation computing resource
Server establishes computing resource information, calculating task information；
Step 2 runs total task schedule process in task scheduling server, transports in every wheel static security analysis function
When the row period arrives, increases new power grid static security analysis in task list and calculate general assignment, and according to static security analysis
The general assignment that calculates of forecast failure resolve into the specific parallel computation task of each computing resource, be written to computing resource table
In, and prepare to calculate data；
Step 3, on computing resource server, computing resource finger daemon is each according to what is updated in concurrent computation resource table
The task of computing resource distributes state, starts corresponding computing resource calculation procedure；
Step 4, in task scheduling server, wait next static security analysis calculating cycle arrive when, return is held
Row step 2.
Moreover, the specific steps of the step 1 include:
(1-1) establishes task list, comprising: task name, task type, task start personnel, task calculating module, task meter
Calculate function, task time-out time, task computation family, task current state, task status variation markers, the task start time,
Task completion time, task complete ratio, task current operational information；
(1-2) establishes computing resource table, comprising: resource name, Resource Server title, resource module title, resource module
Whether function title puts into operation, Resource Calculation module process, Resource Calculation task status, the resource same day frequency of failure, wherein
" Resource Calculation task status " can be with are as follows: " free time ", " publication is completed ", " calculate and complete ", " calculating failure ", "abnormal"；
(1-3) establishes the record of computing resource table, is denoted as set R=according to the number of servers n for participating in parallel computation
{ Ri, i=1 ..., n }, n are the number of servers using technology of configuration, that is, concurrent computation resource quantity；And to each
The resource tasks state of computing resource REi is initially set to " free time "；
(1-4) establishes electric network fault set according to grid condition, is denoted as NF={ NFx, x=1 ..., n2 }, and n2 is maximum
Number of faults；
Moreover, the specific steps of the step 2 include:
(2-1) firstly generates the model file Fz of electric network swim calculating, and file includes the connection relationship of each branch in power grid
And the trend of each node injects data, model file is stored using compress mode；
(2-2) scanography computing resource table finds the resource that all resource tasks states are set to " free time ", is recorded as
Set RE={ REi, i=1 ..., n1 }, total quantity n1；
(2-3) automatically generates n2 calculating subtask: TF={ TFj, j=1 ..., n2 } according to electric network fault set NF,
Distribution method is according to equalitarian distribution method, each calculating number of faults for calculating subtask and undertaking are as follows: n3=n2/n1+1；n3
The calculating number of faults undertaken for each calculating subtask；To generate each assignment file TFile for calculating subtask, it is denoted as
Include interior n3 failure in TFF={ TFFj, j=1 ..., n2 }, each file TFFj, is denoted as TFFj-k, k=1 ..., n3, text
Part uses the text file of E language description, includes resource state information and forecast failure collection information；
(2-4) carries out task distribution to each calculating subtask TFj in set TF；
(2-5) scanography computing resource table R judges the resource tasks state of each resource Ri in table；
Moreover, the specific steps of step 2 (2-3) step include:
(2-3-1) generates the resource state information in TFFj file, class name calsc_info, and the following table of attribute is retouched
It states；Each subtask corresponds to data line in data block calsc_info；
(2-3-2) generates the forecast failure information class name flt_info in TFFj file, the following table description of attribute；Each
Forecast failure TFFj-k corresponds to data line in data block flt_info；
Moreover, the specific steps of (2-4) step of the step 2 include:
(2-4-1) scanography computing resource table R finds the resource that some resource tasks state is set to " free time ", record
For Ri；
Subtask TFj is distributed to computing resource Ri by (2-4-2), and method is: sending tide to the server of computing resource Ri
Flow calculation model file Fz, calculation document TFFj, the resource tasks state of juxtaposition computing resource Ri are set to " publication is completed "；
Moreover, the specific steps of (2-5) step of the step 2 include:
(2-5-1) returns resource Ri calculated result file FRsti if resource tasks state is " calculate and complete "
It receives, i.e., destination file FRsti is transferred in task scheduling server, and the content in each destination file is shown man-machine
On interface；After the completion of destination file FRsti recycling, the resource tasks state for being placed in resource Ri is set to " free time "；
(2-5-2) holds the subtask TFj for distributing to the resource if resource tasks state is " calculating failure " again
(2-4) step of row step 2 carries out calculating task and redistributes；The calculating frequency of failure m on the resource Ri same day is carried out simultaneously tired
Add, when m is less than specified frequency of failure threshold value N, the calculating task state for setting resource Ri is " free time "；When m is more than or equal to N
When, the calculating task state for setting resource Ri is "abnormal", and resource Ri is no longer participate in the same day subsequent calculating task；
Moreover, the specific steps of the step 3 include:
(3-1) in computing resource table, the resource tasks state of the corresponding resource record Ri of current server is sentenced
Disconnected, if the calculating task state of resource is " publication is completed ", startup separator analytical calculation process is calculated, and is walked after execution
Rapid 3 (3-2) step；Otherwise any step is not executed；
(3-2) reads in the model file Fz that newest received electric network swim calculates, and forms power grid computation model；
(3-3) reads in newest received assignment file TFFj, to wherein each failure TFFj-k, executes following steps:
The branch or node failure is arranged in (3-3-1) simulation in electric network model；
(3-3-2) carries out electric network swim calculating, if calculated successfully, executes step (3-3-3), if calculating failure, directly
It meets (3-6) for executing step 3 and terminates calculating；
The out-of-limit situation of branch, node of (3-3-3) according to power grid after calculation of tidal current statistic record failure, forms and calculates
As a result TFFRsti is stored into calculator memory table；
(3-3-4) restores the branch of fault setting and node state is original state；
(3-3-5) return step 3-3-1) until the completion of whole calculation of fault；
After the completion of (3-4) whole calculation of fault, calculated result is generated, the calculated result file of current computing resource Ri is written
In FRsti；
The failure that (3-4-1) generates in calculated result file Frsti causes the more limit information, class name of voltage: falt_v_
Rst, it is as follows comprising attribute information；Each failure causes one of the out-of-limit corresponding falt_v_rst data block of voltage to record；
The failure that (3-4-2) generates in calculated result file Frsti causes the more limit information of branch, class name: falt_q_
Rst, it is as follows comprising attribute information；Each failure causes one of the out-of-limit corresponding falt_q_rst data block of branch to record；
The resource tasks state of the corresponding computing resource Ri of current server is set to " calculate and complete " by (3-5), and is returned
(3-1) of step 3；
It is by the write-in computing resource Ri task status of the corresponding computing resource Ri of current server after (3-6) calculating failure
" calculating failure ", and (3-1) of return step 3.
The advantages of the present invention:
1, a kind of electric network fault parallel parsing calculation method based on exchange files of the invention, parallel computation environment system
It is divided into " man-machine interface ", " task management ", " computing resource " three level compositions.Man-machine interface be responsible for configure computing resource table,
Task list data；" task management " operates on primary server, and " computing resource " can be multiple servers, on every server
" accident analysis " computing function is installed, parallel computation may be implemented in multiple servers.Task management layer and computing resource layer pass through
File transmits interaction data, while task management layer also can receive the message of man-machine interface transmission, responds human-computer interaction.It is right
" accident analysis " module drives calculating task with the task list of network share, using input, output file as data interaction Jie
Matter.This method can shorten " accident analysis " and calculate total duration, and the on-line analysis for improving static security analysis function calculates speed
Degree.
2, the present invention in task scheduling server memory by defining task list, computing resource table, by defining E file
As the data interaction medium of input and output, task scheduling server is unified to be allocated and issues to task, while more meters
Resource Server parallel computation is calculated, greatly shortens and calculates the time；The on-line analysis for improving static security analysis function calculates
Speed meets minute grade calculating demand；And then improve the stabilization of power grid.
Detailed description of the invention
Fig. 1 is the flow chart that task server of the invention executes；
Fig. 2 is the flow chart that the Resource Server of this hair executes.
Specific embodiment
The embodiment of the present invention is described in further detail below in conjunction with attached drawing:
Electric power system data markup language --- E linguistic norm (hereinafter referred to as " E language ") is in IEC61970-301 electricity
On the basis of the object-oriented of Force system common data model CIM (Common Information Model) is abstract, exist for CIM
The new and effective electric power system data markup language of prepared one kind of efficiency defect when being described in a manner of XML.E language
Data file includes annotation (optional), system statement, data block start mark, data block head definition, data block, end of data block
The parts such as label.It is generally divided into class name definition, attribute definition etc.；State Grid Corporation of China's company standard " Q-GDW-215-2008 electricity
Force system data markup language --- E linguistic norm " the specified in more detail describing mode of E language, specific format be defined as follows:
1) it annotates
Annotation is the optional part of E language, is started by double slashes " // " guidance.Annotation may alternatively appear in appointing for data file
Meaning place, both can independent a line, the rear portion for also allowing to be expert at.
2) system is stated
E language system statement be located at data block front end, by left angle brackets add exclamation side by side " <！" guidance start, by exclamation
Add right angle bracket side by side "！> " label terminate.Illustrate the application system title, the E language version of use that generate data in system statement
Originally, character set and versions of data.Particular content in label is as follows:
A) System=application system title.Refer to the application system title for generating notebook data.
B) Version=E language version.Refer to E language version used by encapsulation notebook data.Due to the hair of E language itself
Exhibition, syntax rule
It may change with symbol, this version information can be used for the difference of procedure identification language itself.
C) Code=character set title.Refer to character set used in this document.
D) Data=versions of data.Refer to version locating for notebook data.Since application demand changes, its envelope of the data of same classification
Assembling structure may will do it local expansion, modification or simplification, this version information can be used for the subtle of procedure identification data itself
Difference.
Divided between above each section with space, and can be extended according to system situation.The following are a system sound
Bright practical sample: <！System=SYS Version=1.0 Code=UTF-8 Data=1.0！>
3) data block starting and ending
The starting of E language data block angle brackets "<class name>" or "<class name:: physical name>" indicate that end of data block is used
In angle brackets plus monocline thick stick "</class name >" or "</class name ::entity name >" indicate." class name:: physical name " in, the former number
According to the class of data in block object, the latter indicates the ownership object of data data in block.
4) data block head defines
Data block head definition with address character " " mark, for showing the basic structure of data: i.e. horizontal tabular form, single-column type or
Multi-column type.The specific label that data block head defines is as follows: single-address, which accords with "@", indicates that the basic structure of data is horizontal tabular form, language
Method are as follows: 1 attribute-name of@serial number attribute-name, 2 attribute-name 3 ... "@", serial number are divided using space between each attribute-name, and serial number also can be used
Data ID indicates (similarly hereinafter), and single-column type is similar with processing mode of the multi-column type structure for serial number.The data block of horizontal tabular form structure
In each object accounts for a line, each attribute accounts for a column, be suitable for table class data or attribute is less and object is more data.
5) data block
Data block is made of multiple data lines, and each data line is guided by pound sign " # ".Between each data value in data line
Divided by space.It is corresponded according to capable expression way and data block head definition.The first row of data block is always in " # " heel
One or more spaces, behind space plus serial number, such as " #10 ", making serial number individually becomes a column.The serial number indicates current row data at this
Sequence in data block, is numbered with 1.Serial number is followed by starting data value sequence behind a space.
Text interaction in the present invention is all using E language description, and what definition data block used is all horizontal tabular form, i.e., often
A class is all a table, has different attributes in each table.
A kind of electric network fault parallel parsing calculation method based on exchange files, comprising the following steps:
Step 1, before calculating, establish task list, computing resource table, establish system failure collection；Task list, computing resource table
It establishes in the memory of server, other Resource Servers can be by network interface access task list and computing resource table
Data；Computing resource information, calculating task information are established according to computing resource server is participated in；
The specific steps of the step 1 include:
(1-1) establishes task list (DEF_TASK), includes: task name, is appointed at task type (automatic to calculate, artificial to calculate)
Be engaged in starting personnel, task calculating module, task computation function, task time-out time, task computation family, task current state,
Task status changes markers, task start time, task completion time, task completion ratio, task current operational information；
(1-2) establishes computing resource table (DEF_CALSC), comprising: resource name, Resource Server title, resource module name
Claim, resource module function title, whether put into operation, Resource Calculation module process, Resource Calculation task status, the mistake of the resource same day
Number is lost, wherein " Resource Calculation task status " can be with are as follows: " free time ", " publication is completed ", " calculate and complete ", " calculating failure ",
"abnormal"；
(1-3) establishes the record of computing resource table, is denoted as set R=according to the number of servers n for participating in parallel computation
{ Ri, i=1 ..., n }, n are the number of servers using technology of configuration, that is, concurrent computation resource quantity；And to each
The resource tasks state of computing resource REi is initially set to " free time "；
(1-4) establishes electric network fault set according to grid condition, is denoted as NF={ NFx, x=1 ..., n2 }, and n2 is maximum
Number of faults；
Step 2 runs total task schedule process in task scheduling server, transports in every wheel static security analysis function
When the row period arrives, increases new power grid static security analysis in task list and calculate general assignment, and according to static security analysis
The general assignment that calculates of forecast failure resolve into the specific parallel computation task of each computing resource, be written to computing resource table
In, and prepare to calculate data；
As shown in Figure 1, the specific steps of the step 2 include:
(2-1) firstly generates the model file Fz of electric network swim calculating, and file includes the connection relationship of each branch in power grid
And the trend of each node injects data, model file is stored using compress mode；
(2-2) scanography computing resource table finds the resource that all resource tasks states are set to " free time ", is recorded as
Set RE={ REi, i=1 ..., n1 }, total quantity n1；
(2-3) automatically generates n2 calculating subtask: TF={ TFj, j=1 ..., n2 } according to electric network fault set NF,
Distribution method is according to equalitarian distribution method, each calculating number of faults for calculating subtask and undertaking are as follows: n3=n2/n1+1；n3
The calculating number of faults undertaken for each calculating subtask；To generate each assignment file TFile for calculating subtask, it is denoted as
Include interior n3 failure in TFF={ TFFj, j=1 ..., n2 }, each file TFFj, is denoted as TFFj-k, k=1 ..., n3, text
Part uses the text file of E language description, includes resource state information and forecast failure collection information；
The specific steps of step 2 (2-3) step include:
(2-3-1) generates the resource state information in TFFj file, class name (calsc_info), and the following table of attribute is retouched
It states；Each subtask corresponds to data line in data block calsc_info；
| Serial number | Attribute | Data type | Explanation |
| 1 | Resource module title | Character string | |
| 2 | Resource status | Shaping | |
| 3 | Current calculating task ID | Long shaping | |
| 4 | Current calculating task function | Character string | |
| 5 | Current calculating task starts the time | Shaping | |
| 6 | The current calculating task end time | Shaping | |
| 7 | Current calculating task completes ratio | Floating type |
(2-3-2) generates the forecast failure information class name (flt_info) in TFFj file, the following table description of attribute；Often
A forecast failure TFFj-k corresponds to data line in data block flt_info；
(2-4) carries out task distribution to each calculating subtask TFj in set TF；
The specific steps of (2-4) step of the step 2 include:
(2-4-1) scanography computing resource table R finds the resource that some resource tasks state is set to " free time ", record
For Ri；
Subtask TFj is distributed to computing resource Ri by (2-4-2), and method is: sending tide to the server of computing resource Ri
Flow calculation model file Fz, calculation document TFFj, the resource tasks state of juxtaposition computing resource Ri are set to " publication is completed "；
(2-5) scanography computing resource table R judges the resource tasks state of each resource Ri in table；
The specific steps of (2-5) step of the step 2 include:
(2-5-1) returns resource Ri calculated result file FRsti if resource tasks state is " calculate and complete "
It receives, i.e., destination file FRsti is transferred in task scheduling server, and the content in each destination file is shown man-machine
On interface (destination file format is referring to step 3))；After the completion of destination file FRsti recycling, it is placed in the resource tasks of resource Ri
State is set to " free time "；
(2-5-2) holds the subtask TFj for distributing to the resource if resource tasks state is " calculating failure " again
(2-4) step of row step 2 carries out calculating task and redistributes；The calculating frequency of failure m on the resource Ri same day is carried out simultaneously tired
Add, when m is less than specified frequency of failure threshold value N, the calculating task state for setting resource Ri is " free time "；When m is more than or equal to N
When, the calculating task state for setting resource Ri is "abnormal", and resource Ri is no longer participate in the same day subsequent calculating task；
Step 3, on computing resource server, computing resource finger daemon is each according to what is updated in concurrent computation resource table
The task of computing resource distributes state, starts corresponding computing resource calculation procedure.
As shown in Fig. 2, the specific steps of the step 3 include:
(3-1) in computing resource table, the resource tasks state of the corresponding resource record Ri of current server is sentenced
Disconnected, if the calculating task state of resource is " publication is completed ", startup separator analytical calculation process is calculated, and is walked after execution
Rapid 3 (3-2) step；Otherwise any step is not executed；
(3-2) reads in the model file Fz that newest received electric network swim calculates, and forms power grid computation model；
(3-3) reads in newest received assignment file TFFj, to wherein each failure TFFj-k, executes following steps:
The branch or node failure is arranged in (3-3-1) simulation in electric network model；
(3-3-2) carries out electric network swim calculating, if calculated successfully, executes step (3-3-3), if calculating failure, directly
It meets (3-6) for executing step 3 and terminates calculating；
The out-of-limit situation of branch, node of (3-3-3) according to power grid after calculation of tidal current statistic record failure, forms and calculates
As a result TFFRsti is stored into calculator memory table；
(3-3-4) restores the branch of fault setting and node state is original state；
(3-3-5) return step 3-3-1) until the completion of whole calculation of fault.
After the completion of (3-4) whole calculation of fault, calculated result is generated, the calculated result file of current computing resource Ri is written
In FRsti；
The failure that (3-4-1) generates in calculated result file Frsti causes the more limit information, class name of voltage: falt_v_
Rst, it is as follows comprising attribute information；Each failure causes one of the out-of-limit corresponding falt_v_rst data block of voltage to record；
| Serial number | Attribute | Data type | Explanation |
| 1 | Failure ID | Long shaping | |
| 2 | Failure title | Character string | |
| 3 | Failure causes more upper limit bus quantity | Shaping | |
| 4 | Highest gets over upper limit bus title | Character string | |
| 5 | Highest gets over upper limit bus voltage value | Floating type | |
| 6 | Failure causes more lower limit bus quantity | Shaping | |
| 7 | Minimum more lower limit bus title | Character string | |
| 8 | Minimum more lower limit bus voltage value | Floating type |
The failure that (3-4-2) generates in calculated result file Frsti causes the more limit information of branch, class name: falt_q_
Rst, it is as follows comprising attribute information；Each failure causes one of the out-of-limit corresponding falt_q_rst data block of branch to record.
| Serial number | Attribute | Data type | Explanation |
| 1 | Failure ID | Long shaping | |
| 2 | Failure title | Character string | |
| 3 | Failure causes the out-of-limit quantity of branch | Shaping | |
| 4 | Highest gets over upper limit branch title | Character string | |
| 5 | Highest, which gets over upper limit Branch Power Flow, work value | Floating type | |
| 6 | Highest gets over upper limit Branch Power Flow without work value | Floating type |
The resource tasks state of the corresponding computing resource Ri of current server is set to " calculate and complete " by (3-5), and is returned
(3-1) of step 3；
It is by the write-in computing resource Ri task status of the corresponding computing resource Ri of current server after (3-6) calculating failure
" calculating failure ", and (3-1) of return step 3；
Step 4, in task scheduling server, wait next static security analysis calculating cycle arrive when, return is held
Row step 2.
The working principle of the invention is:
A kind of electric network fault parallel parsing calculation method based on exchange files proposed by the present invention, the present invention be by
Task list, computing resource table are defined in task scheduling server memory, by defining E file as the data interaction of input and output
Medium, task scheduling server is unified to be allocated and issues to task, while more computing resource server parallel computations, greatly
Big shortening calculates the time；The on-line analysis calculating speed for improving static security analysis function, meeting minute grade calculating needs
It asks；And then improve the stabilization of power grid.
It is emphasized that embodiment of the present invention be it is illustrative, without being restrictive, therefore the present invention includes
It is not limited to embodiment described in specific embodiment, it is all to be obtained according to the technique and scheme of the present invention by those skilled in the art
Other embodiments, also belong to the scope of protection of the invention.
Claims (7)
1. a kind of electric network fault parallel parsing calculation method based on exchange files, it is characterised in that: the following steps are included:
Step 1, before calculating, establish task list, computing resource table, establish system failure collection；According to participation computing resource service
Device establishes computing resource information, calculating task information；
Step 2 runs total task schedule process in task scheduling server, runs week in every wheel static security analysis function
When phase arrives, increases new power grid static security analysis in task list and calculate general assignment, and according to the pre- of static security analysis
Think that the general assignment of calculation of fault resolves into the specific parallel computation task of each computing resource, be written in computing resource table,
And prepare to calculate data；
Step 3, on computing resource server, computing resource finger daemon, and each according to what is updated in concurrent computation resource table
The task of computing resource distributes state, starts corresponding computing resource calculation procedure；
Step 4, in task scheduling server, wait next static security analysis calculating cycle arrive when, return execute step
Rapid 2.
2. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 1, feature exist
In: the specific steps of the step 1 include:
(1-1) establishes task list, comprising: task name, task type, task start personnel, task calculating module, task computation function
Energy, task time-out time, task computation family, task current state, task status change markers, task start time, task
Deadline, task complete ratio, task current operational information；
(1-2) establishes computing resource table, comprising: resource name, Resource Server title, resource module title, resource module function
Whether title puts into operation, Resource Calculation module process, Resource Calculation task status, the resource same day frequency of failure, wherein " money
Source calculating task state " can be with are as follows: " free time ", " publication is completed ", " calculate and complete ", " calculating failure ", "abnormal"；
(1-3) establishes the record of computing resource table, is denoted as set R={ Ri, i according to the number of servers n for participating in parallel computation
=1 ..., n }, n is the number of servers using technology of configuration, that is, concurrent computation resource quantity；And each calculating is provided
The resource tasks state of source REi is initially set to " free time "；
(1-4) establishes electric network fault set according to grid condition, is denoted as NF={ NFx, x=1 ..., n2 }, and n2 is maximum failure
Number.
3. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 1, feature exist
In: the specific steps of the step 2 include:
(2-1) firstly generate electric network swim calculating model file Fz, file include in power grid the connection relationship of each branch and
The trend of each node injects data, and model file is stored using compress mode；
(2-2) scanography computing resource table finds the resource that all resource tasks states are set to " free time ", is recorded as gathering
RE={ REi, i=1 ..., n1 }, total quantity n1；
(2-3) automatically generates n2 calculating subtask: TF={ TFj, j=1 ..., n2 } according to electric network fault set NF, distribution
Method is according to equalitarian distribution method, each calculating number of faults for calculating subtask and undertaking are as follows: n3=n2/n1+1；N3 is every
A calculating number of faults for calculating subtask and undertaking；To generate each assignment file TFile for calculating subtask, it is denoted as TFF=
{ TFFj, j=1 ..., n2 } includes interior n3 failure in each file TFFj, is denoted as TFFj-k, k=1 ..., n3, file use
The text file of E language description includes resource state information and forecast failure collection information；
(2-4) carries out task distribution to each calculating subtask TFj in set TF；
(2-5) scanography computing resource table R judges the resource tasks state of each resource Ri in table.
4. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 3, feature exist
In: the specific steps of step 2 (2-3) step include:
(2-3-1) generates the resource state information in TFFj file, class name calsc_info, the following table description of attribute；Often
A subtask corresponds to data line in data block calsc_info；
(2-3-2) generates the forecast failure information class name flt_info in TFFj file, the following table description of attribute；Each anticipation
Failure TFFj-k corresponds to data line in data block flt_info.
5. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 3, feature exist
In: the specific steps of (2-4) step of the step 2 include:
(2-4-1) scanography computing resource table R finds the resource that some resource tasks state is set to " free time ", is recorded as Ri；
Subtask TFj is distributed to computing resource Ri by (2-4-2), and method is: sending trend meter to the server of computing resource Ri
Model file Fz, calculation document TFFj are calculated, the resource tasks state of juxtaposition computing resource Ri is set to " publication is completed ".
6. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 3, feature exist
In: the specific steps of (2-5) step of the step 2 include:
(2-5-1) recycles resource Ri calculated result file FRsti if resource tasks state is " calculate and complete ",
Destination file FRsti is transferred in task scheduling server, and the content in each destination file is shown on man-machine boundary
On face；After the completion of destination file FRsti recycling, the resource tasks state for being placed in resource Ri is set to " free time "；
(2-5-2) re-executes step to the subtask TFj for distributing to the resource if resource tasks state is " calculating failure "
Rapid 2 (2-4) step carries out calculating task and redistributes；The calculating frequency of failure m on the resource Ri same day is added up simultaneously,
When m is less than specified frequency of failure threshold value N, the calculating task state for setting resource Ri is " free time "；When m is more than or equal to N,
The calculating task state for setting resource Ri is "abnormal", and resource Ri is no longer participate in the same day subsequent calculating task.
7. a kind of electric network fault parallel parsing calculation method based on exchange files according to claim 1, feature exist
In: the specific steps of the step 3 include:
(3-1) in computing resource table, the resource tasks state of the corresponding resource record Ri of current server is judged, such as
The calculating task state of fruit resource is " publication is completed ", then startup separator analytical calculation process is calculated, step 3 after execution
(3-2) step；Otherwise any step is not executed；
(3-2) reads in the model file Fz that newest received electric network swim calculates, and forms power grid computation model；
(3-3) reads in newest received assignment file TFFj, to wherein each failure TFFj-k, executes following steps:
The branch or node failure is arranged in (3-3-1) simulation in electric network model；
(3-3-2) carries out electric network swim calculating, if calculated successfully, executes step (3-3-3), if calculating failure, directly holds
(3-6) of row step 3 simultaneously terminates calculating；
The out-of-limit situation of branch, node of (3-3-3) according to power grid after calculation of tidal current statistic record failure, forms calculated result
TFFRsti is stored into calculator memory table；
(3-3-4) restores the branch of fault setting and node state is original state；
(3-3-5) return step (3-3-1) is completed until whole calculation of fault；
After the completion of (3-4) whole calculation of fault, calculated result is generated, the calculated result file of current computing resource Ri is written
In FRsti；
The failure that (3-4-1) generates in calculated result file Frsti causes the more limit information, class name of voltage: falt_v_rst, packet
It is as follows containing attribute information；Each failure causes one of the out-of-limit corresponding falt_v_rst data block of voltage to record；
The failure that (3-4-2) generates in calculated result file Frsti causes the more limit information of branch, class name: falt_q_rst, packet
It is as follows containing attribute information；Each failure causes one of the out-of-limit corresponding falt_q_rst data block of branch to record；
The resource tasks state of the corresponding computing resource Ri of current server is set to " calculate and complete " by (3-5), and return step 3
(3-1)；
(3-6) calculate failure after by the write-in computing resource Ri task status of the corresponding computing resource Ri of current server be " meter
Calculate failure ", and (3-1) of return step 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811166572.1A CN109167354B (en) | 2018-10-08 | 2018-10-08 | Power grid expected fault parallel analysis and calculation method based on file exchange |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811166572.1A CN109167354B (en) | 2018-10-08 | 2018-10-08 | Power grid expected fault parallel analysis and calculation method based on file exchange |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109167354A true CN109167354A (en) | 2019-01-08 |
| CN109167354B CN109167354B (en) | 2022-02-22 |
Family
ID=64877414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811166572.1A Active CN109167354B (en) | 2018-10-08 | 2018-10-08 | Power grid expected fault parallel analysis and calculation method based on file exchange |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109167354B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117955985A (en) * | 2024-03-26 | 2024-04-30 | 佛山科学技术学院 | Node pressure optimizing method based on parallel computing |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101441580A (en) * | 2008-12-09 | 2009-05-27 | 华北电网有限公司 | Distributed paralleling calculation platform system and calculation task allocating method thereof |
| CN101685480A (en) * | 2008-09-27 | 2010-03-31 | 国家电力调度通信中心 | Parallel computing method for security and stability analysis of large power grid and computing platform |
| CN103325012A (en) * | 2013-06-18 | 2013-09-25 | 国家电网公司 | Parallel computing dynamic task distribution method applicable to grid security correction |
| CN103873321A (en) * | 2014-03-05 | 2014-06-18 | 国家电网公司 | Distributed file system-based simulation distributed parallel computing platform and method |
| CN104239144A (en) * | 2014-09-22 | 2014-12-24 | 珠海许继芝电网自动化有限公司 | Multilevel distributed task processing system |
| CN104993477A (en) * | 2015-06-29 | 2015-10-21 | 中国电力科学研究院 | Static safety parallel analysis method for AC-DC power grid based on configuration mode |
| CN105576648A (en) * | 2015-11-23 | 2016-05-11 | 中国电力科学研究院 | Static security analysis double-layer parallel method based on GPU-CUP heterogeneous computing platform |
| CN105610157A (en) * | 2016-01-27 | 2016-05-25 | 国网山东省电力公司烟台供电公司 | Fault-driven static security analysis system of regional power grid |
| CN106058856A (en) * | 2016-07-01 | 2016-10-26 | 中国南方电网有限责任公司 | Method for analyzing static security of grid rapidly |
| JP2016224832A (en) * | 2015-06-03 | 2016-12-28 | 富士通株式会社 | Dump management device, dump management program, and dump management method |
| CN106874113A (en) * | 2017-01-19 | 2017-06-20 | 国电南瑞科技股份有限公司 | A kind of many GPU heterogeneous schemas static security analysis computational methods of CPU+ |
| CN107132762A (en) * | 2017-04-17 | 2017-09-05 | 国电南瑞科技股份有限公司 | Online Static security assessment method based on forecast failure collection automatic screening |
| US20170279703A1 (en) * | 2016-03-25 | 2017-09-28 | Advanced Micro Devices, Inc. | Managing variations among nodes in parallel system frameworks |
| CN108256263A (en) * | 2018-02-07 | 2018-07-06 | 中国电力科学研究院有限公司 | A kind of electric system hybrid simulation concurrent computational system and its method for scheduling task |
| CN108521122A (en) * | 2018-04-13 | 2018-09-11 | 深圳供电局有限公司 | A kind of Contingency Analysis of Power Systems method considering prepared auto restart |
-
2018
- 2018-10-08 CN CN201811166572.1A patent/CN109167354B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101685480A (en) * | 2008-09-27 | 2010-03-31 | 国家电力调度通信中心 | Parallel computing method for security and stability analysis of large power grid and computing platform |
| CN101441580A (en) * | 2008-12-09 | 2009-05-27 | 华北电网有限公司 | Distributed paralleling calculation platform system and calculation task allocating method thereof |
| CN103325012A (en) * | 2013-06-18 | 2013-09-25 | 国家电网公司 | Parallel computing dynamic task distribution method applicable to grid security correction |
| CN103873321A (en) * | 2014-03-05 | 2014-06-18 | 国家电网公司 | Distributed file system-based simulation distributed parallel computing platform and method |
| CN104239144A (en) * | 2014-09-22 | 2014-12-24 | 珠海许继芝电网自动化有限公司 | Multilevel distributed task processing system |
| JP2016224832A (en) * | 2015-06-03 | 2016-12-28 | 富士通株式会社 | Dump management device, dump management program, and dump management method |
| CN104993477A (en) * | 2015-06-29 | 2015-10-21 | 中国电力科学研究院 | Static safety parallel analysis method for AC-DC power grid based on configuration mode |
| CN105576648A (en) * | 2015-11-23 | 2016-05-11 | 中国电力科学研究院 | Static security analysis double-layer parallel method based on GPU-CUP heterogeneous computing platform |
| CN105610157A (en) * | 2016-01-27 | 2016-05-25 | 国网山东省电力公司烟台供电公司 | Fault-driven static security analysis system of regional power grid |
| US20170279703A1 (en) * | 2016-03-25 | 2017-09-28 | Advanced Micro Devices, Inc. | Managing variations among nodes in parallel system frameworks |
| CN106058856A (en) * | 2016-07-01 | 2016-10-26 | 中国南方电网有限责任公司 | Method for analyzing static security of grid rapidly |
| CN106874113A (en) * | 2017-01-19 | 2017-06-20 | 国电南瑞科技股份有限公司 | A kind of many GPU heterogeneous schemas static security analysis computational methods of CPU+ |
| CN107132762A (en) * | 2017-04-17 | 2017-09-05 | 国电南瑞科技股份有限公司 | Online Static security assessment method based on forecast failure collection automatic screening |
| CN108256263A (en) * | 2018-02-07 | 2018-07-06 | 中国电力科学研究院有限公司 | A kind of electric system hybrid simulation concurrent computational system and its method for scheduling task |
| CN108521122A (en) * | 2018-04-13 | 2018-09-11 | 深圳供电局有限公司 | A kind of Contingency Analysis of Power Systems method considering prepared auto restart |
Non-Patent Citations (4)
| Title |
|---|
| VAHID JALILI-MARANDI 等: "Large-Scale Transient Stability Simulation of Electrical Power Systems on Parallel GPUs", 《IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS》 * |
| 宗永升，张祎: "电力系统在线安全分析并行分布式计算", 《计算机工程与设计》 * |
| 李虎成 等: "基于多进程并行计算的日前调度计划安全校核技术", 《电网与清洁能源》 * |
| 陆娟娟 等: "基于GPU并行的静态安全分析设计与应用", 《计算机科学与探索》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117955985A (en) * | 2024-03-26 | 2024-04-30 | 佛山科学技术学院 | Node pressure optimizing method based on parallel computing |
| CN117955985B (en) * | 2024-03-26 | 2024-06-07 | 佛山科学技术学院 | Node pressure optimizing method based on parallel computing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109167354B (en) | 2022-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107292003B (en) | Automatic generation method of power grid station wiring diagram | |
| CN108736465B (en) | Future-state power grid model construction method and device | |
| CN101937435B (en) | Online synchronized method of distributed real-time database confronting power system | |
| CN101996271B (en) | Software interface for automatically generating simulation calculation model of PSCAD power system | |
| CN106844198A (en) | Distributed dispatching automation test platform and method | |
| CN105787089A (en) | Power distribution network scheduling foundation data integrating method | |
| CN112260331B (en) | Extra-high voltage alternating current-direct current power grid simulation platform and construction method | |
| CN104133443A (en) | Method for automatically generating power grid dispatching operation order through power grid overhaul application forms | |
| CN108229799B (en) | Multi-source heterogeneous power grid operation real-time data access system and method | |
| CN108462175A (en) | A kind of electric heating equipment demand response interactive approach, system and device | |
| CN104503827B (en) | A kind of large-scale electrical power system method of operation power flow transfer is than multi-core parallel concurrent batch processing method | |
| CN105373008A (en) | Full physics model simulation method for electric power system | |
| CN116010398A (en) | Method and system for constructing power load modeling data based on station line-to-household relationship | |
| CN107391617A (en) | Model method is led based on monitoring system automatically | |
| CN109167354A (en) | A kind of power grid forecast failure parallel parsing calculation method based on exchange files | |
| CN107908731B (en) | PSD-BPA-based Guangxi power grid area load data batch modification method | |
| CN112541011A (en) | Power grid terminal fault processing method and device based on RPA and AI | |
| CN111711651A (en) | New energy data sharing method and system | |
| CN102393820A (en) | Logical implementation method for controlling building monitoring system | |
| CN112532432B (en) | Automatic mapping method and system based on LTU terminal | |
| CN109656589A (en) | A kind of secondary equipment of intelligent converting station software harnesses method and system | |
| CN108599233A (en) | A kind of power distribution network receiving capacity determining methods and device containing distributed generation resource | |
| CN111090710B (en) | Distribution network multi-time scale graph-model abnormal automatic maintenance method | |
| CN104091225A (en) | Object-oriented electric power system scheduling operation ticket system | |
| CN108898307B (en) | Method for generating production plan on large scale based on strategy by power station equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP02 | Change in the address of a patent holder | ||
| CP02 | Change in the address of a patent holder |
Address after: No. 8, Haitai Huake 4th Road, Huayuan Industrial Zone, High tech Zone, Binhai New Area, Tianjin, 300384 Patentee after: ELECTRIC POWER SCIENCE & RESEARCH INSTITUTE OF STATE GRID TIANJIN ELECTRIC POWER Co. Patentee after: STATE GRID TIANJIN ELECTRIC POWER Co. Patentee after: STATE GRID CORPORATION OF CHINA Patentee after: BEIJING KING STAR HI-TECH SYSTEM CONTROL Co.,Ltd. Address before: No.8, Haitai Huake 4th Road, Xiqing District, Tianjin 300384 Patentee before: ELECTRIC POWER SCIENCE & RESEARCH INSTITUTE OF STATE GRID TIANJIN ELECTRIC POWER Co. Patentee before: STATE GRID TIANJIN ELECTRIC POWER Co. Patentee before: STATE GRID CORPORATION OF CHINA Patentee before: BEIJING KING STAR HI-TECH SYSTEM CONTROL Co.,Ltd. |