CN108520472A - A kind of method, apparatus and electronic equipment of processing electric power system data - Google Patents
A kind of method, apparatus and electronic equipment of processing electric power system data Download PDFInfo
- Publication number
- CN108520472A CN108520472A CN201810166848.XA CN201810166848A CN108520472A CN 108520472 A CN108520472 A CN 108520472A CN 201810166848 A CN201810166848 A CN 201810166848A CN 108520472 A CN108520472 A CN 108520472A
- Authority
- CN
- China
- Prior art keywords
- learning model
- data
- training
- model
- electric
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012545 processing Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000004458 analytical method Methods 0.000 claims description 26
- 238000013136 deep learning model Methods 0.000 claims description 25
- 238000004891 communication Methods 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 13
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 5
- 239000012141 concentrate Substances 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 5
- 238000003012 network analysis Methods 0.000 abstract description 11
- 230000009471 action Effects 0.000 description 12
- 238000003860 storage Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 3
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 3
- 235000003140 Panax quinquefolius Nutrition 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 235000013399 edible fruits Nutrition 0.000 description 3
- 230000005611 electricity Effects 0.000 description 3
- 235000008434 ginseng Nutrition 0.000 description 3
- 230000001174 ascending effect Effects 0.000 description 2
- 238000011217 control strategy Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 239000004020 conductor Substances 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
An embodiment of the present invention provides it is a kind of processing electric power system data method, apparatus and electronic equipment, the method includes：Based on deeply learning model trained in advance, the learning value information in Operation of Electric Systems data is obtained；The operating status of the electric system is judged according to acquired learning value information；According to the operating status of the electric system, control decision is determined, wherein the control decision is for ensureing the power system security stable operation.The technical solution provided through the embodiment of the present invention, can directly it start with from electric network data, the learning value information in Operation of Electric Systems data is obtained based on trained deeply learning model in advance, the operating status of electric system is judged according to acquired learning value information and provides reasonable control program；So as to avoid dependence of the electrical network analysis to physical model, and model framework need not be readjusted for different running method different scenes, training effectiveness is high, with strong applicability.
Description
Technical field
The present invention relates to technical field of data processing, more particularly to a kind of method, apparatus of processing electric power system data
And electronic equipment.
Background technology
Increase in power grid accounting for electricity consumption as new energy, electric vehicle etc. are novel, traditional Power System Analysis method
It is poor for electrical network analysis control applicability under new development situation, power network development requirement cannot be met, specifically, conventional electric power
Network analysis control is often using building math equation group based on physical model modeling and solving, according to result of calculation decision-making system
State.But as NETWORK STRUCTURE PRESERVING POWER SYSTEM is increasingly sophisticated, the analysis method based on physical model is difficult to adapt to power grid demand, point
Credible result degree is analysed to decline.
Invention content
The embodiment of the present invention is designed to provide a kind of method, apparatus and electronic equipment of processing electric power system data,
Dependence to avoid electrical network analysis to physical model ensure that the confidence level of analysis result, and for different running method difference
Scene need not readjust model framework, and training effectiveness is high, strong applicability.Specific technical solution is as follows：
In a first aspect, an embodiment of the present invention provides a kind of method of processing electric power system data, the method includes：
Based on deeply learning model trained in advance, the learning value information in Operation of Electric Systems data is obtained；
The operating status of the electric system is judged according to acquired learning value information；
According to the operating status of the electric system, control decision is determined, wherein the control decision is described for ensureing
Power system security stable operation.
Optionally, the training process of the deeply learning model is as follows：
Deeply learning model is initialized, the deeply learning model is by deep learning model and intensified learning mould
Type is constituted；
Obtain electric network data；
Pretreatment is carried out to the electric network data obtained and forms sample data set；
It is concentrated from the sample data and chooses training data；
Utilize the feature of training data described in deep learning model extraction；
The feature extracted is analyzed using intensified learning model, obtains analysis result, and tie based on the analysis
Fruit judges whether to complete training；
If it is judged that complete training, the deeply learning model of training completion is obtained；
If it is judged that not complete training, the model ginseng of the deep learning model and intensified learning model is adjusted
Number returns to execute from the sample data and concentrates the step of choosing training data.
Optionally, the deep learning model is：Depth convolutional neural networks.
Optionally, the deep learning model is for improving initial data value density and increasing model generalization ability
Model.
Optionally, the intensified learning model is Q-Learning analysis models.
Second aspect, an embodiment of the present invention provides a kind of device of processing electric power system data, described device includes：
Data obtaining module, for based on deeply learning model trained in advance, obtaining Operation of Electric Systems data
In learning value information；
Operating status judgment module, the operation shape for judging the electric system according to acquired learning value information
State；
Control decision determining module determines control decision, wherein institute for the operating status according to the electric system
Control decision is stated for ensureing the power system security stable operation.
Optionally, the training process of the deeply learning model is as follows：
Deeply learning model is initialized, the deeply learning model is by deep learning model and intensified learning mould
Type is constituted；
Obtain electric network data；
Pretreatment is carried out to the electric network data obtained and forms sample data set；
It is concentrated from the sample data and chooses training data；
Utilize the feature of training data described in deep learning model extraction；
The feature extracted is analyzed using intensified learning model, obtains analysis result, and tie based on the analysis
Fruit judges whether to complete training；
If it is judged that complete training, the deeply learning model of training completion is obtained；
If it is judged that not complete training, the model ginseng of the deep learning model and intensified learning model is adjusted
Number returns to execute from the sample data and concentrates the step of choosing training data.
Optionally, the deep learning model is：Depth convolutional neural networks.
Optionally, the deep learning model is for improving initial data value density and increasing model generalization ability
Model.
Optionally, the intensified learning model is Q-Learning analysis models.
The third aspect, an embodiment of the present invention provides a kind of electronic equipment, including processor, communication interface, memory and
Communication bus, wherein processor, communication interface, memory complete mutual communication by communication bus；
Memory, for storing computer program；
Processor when for executing the program stored on memory, realizes any processing electric power described in first aspect
The method of system data.
Fourth aspect, an embodiment of the present invention provides a kind of computer readable storage medium, the computer-readable storage
Dielectric memory contains computer program, and the computer program realizes any described in above-mentioned first aspect when being executed by processor
The method for handling electric power system data.
Compared with prior art, technical solution provided in an embodiment of the present invention can directly start with from electric network data, be based on
Trained deeply learning model obtains the learning value information in Operation of Electric Systems data in advance, then according to being obtained
The learning value information taken judges the operating status of electric system and provides reasonable control program；So as to a certain extent
Dependence of the electrical network analysis to physical model is avoided, and model support need not be readjusted for different running method different scenes
Structure, training effectiveness is high, with strong applicability；In addition, the embodiment of the present invention makes full use of data resource, it ensure that resource is not unrestrained
Take.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.
A kind of flow chart of the method for processing electric power system data that Fig. 1 is provided by the embodiment of the present invention；
The depth convolutional neural networks Organization Chart that Fig. 2 is provided by the embodiment of the present invention；
The 16-17 branch trouble differences that Fig. 3 is provided by the embodiment of the present invention cut machine strategy generator deviation comparison diagram；
The comparison different running method incision machine result of decision schematic diagram that Fig. 4 is provided by the embodiment of the present invention；
A kind of structural schematic diagram of the device for processing electric power system data that Fig. 5 is provided by the embodiment of the present invention；
The structural schematic diagram for a kind of electronic equipment that Fig. 6 is provided by the embodiment of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
In order to solve technical problem of the existing technology, an embodiment of the present invention provides a kind of processing electric power system datas
Method, apparatus and electronic equipment, the dependence to avoid electrical network analysis to physical model, and for different running method difference field
Scape need not readjust model framework, and training effectiveness is high, strong applicability.
In a first aspect, the method for being provided for the embodiments of the invention processing electric power system data first below is situated between
It continues.
As shown in Figure 1, a kind of method for processing electric power system data that the embodiment of the present invention is provided, including walk as follows
Suddenly：
S101 obtains the learning value in Operation of Electric Systems data based on deeply learning model trained in advance
Information；
Traditional Power System Analysis control is often used based on physical model modeling structure math equation group and is solved, root
According to result of calculation decision-making system state.As NETWORK STRUCTURE PRESERVING POWER SYSTEM is increasingly sophisticated, the analysis method based on physical model is difficult to fit
Power grid demand, analysis result confidence level is answered to decline.In addition, due in power grid uncertain factor increase, physical model is difficult to pair
Uncertain factor models, and causes power network modeling difficult, analysis cannot be calculated for some problems.
In order to solve traditional the problems of Power System Analysis method, deeply provided in an embodiment of the present invention
Learning model is directly started with from electric network data, learns value information in data, is analyzed operation of power networks state and is provided reasonable control
Scheme processed.The present invention can avoid dependence of the electrical network analysis to physical model to a certain extent, with strong applicability, for different fortune
Line mode different scenes need not readjust model framework, and training effectiveness is high.
Wherein, above-mentioned deeply learning model is made of deep learning and intensified learning two parts.The embodiment of the present invention
In deep learning use depth convolutional neural networks；Intensified learning uses Q-Learning analysis methods.Depth convolutional Neural
Network acquires different characteristic using different convolution kernels, can promote initial data value density, increases model generalization ability.
It should be noted that the Organization Chart of above-mentioned depth convolutional neural networks is as shown in Fig. 2, convolutional neural networks pass through not
Same convolution kernel analysis different characteristic, each convolution kernel capture feature in such a way that weights are shared.First time convolution uses 4
Convolution kernel analysis sample data, generates 4 kinds of feature samples, each sample characteristics weights is identical.Double sampling process is pond
Process can effectively reduce feature samples dimension, while ensure that key message is not lost as far as possible.
Above-mentioned intensified learning Q-Learning methods mainly handle analysis power grid action message, and power grid action effect is passed through
Q functions are quantified as numerical index Q values, and control strategy is obtained by comparing Q value sizes acts of determination quality.
S102 judges the operating status of the electric system according to acquired learning value information；
Acquired learning value information can judge the operating status of electric system, specifically, deeply learns mould
The flow of type training is：Input electric network data；Pretreatment is carried out to the electric network data of input and forms sample data set；Choose training
Data；The feature of training data is extracted using convolutional neural networks；The feature extracted is analyzed using intensified learning, and
Judgement is enough completion training；If it is judged that complete training, then stores parameter and instructed to get to deeply learning model
Practice；If it is judged that not complete training, then the step of choosing training pattern is continued to execute.
After obtaining deeply learning model, when extracting training sample data, every time while a plurality of data, packet are trained
Electric network data under different running method is included, to improve model generalization ability, while improving training effectiveness.Deeply learnt
Cheng Zhong, Reward Program have important role for modelling effect.The variance of data can characterize physical system energy, random matrix reason
The return value after data variance is executed as action is calculated by central-limit theorem.Consider the front and back physical characteristic of power grid action,
The data source calculated as return value using alternator speed deviation data.When power grid normal operation, alternator speed deviation is equal
Even to be distributed near zero, desired value is approximately zero.Power grid, which is in a state of emergency and takes, cuts motor-driven work, and action effectively then generates electricity
Machine velocity deviation reduces, and data fluctuations energy also accordingly reduces.In order to avoid Q values overestimate to result of calculation bring not really
It is qualitative, in conjunction with double Q network and competition Q network struction algorithm models.The object function of the double Q network of intensified learning is (1)
By formula (1), return value more little trick is more effective, i.e., control effect is better.Competition Q networks consideration divides return value
Power grid environment data are divided into running environment data and are moved in conjunction with operation of power networks characteristic for environment return value and action return value
Make information data.Running environment information mainly include generator power information, action message data mainly include voltage, generator rotor angle and
The information such as alternator speed deviation.Based on described above, the variance of alternator speed deviation can reflect control to a certain extent
The quality of strategy, therefore using the variance of alternator speed deviation as return, formula (14) is used as Reward Program.Intensified learning process
In, Q-Learning networks are by continuous corrective networks parameter so that the Q values of network calculations are constantly close to return value.
Wherein, sample data considers that generator reactive, generator electromagnetic work is respectively adopted in power grid various dimensions and physical features
The attribute dimensions such as rate, generator mechanical power, node voltage, alternator speed deviation and generator's power and angle.It is explained for ease of analysis
State, using IEEE39 node systems emulation data as sample data, will cut it is motor-driven number, generator node is ascending,
It is 1-30 that the ascending sequence of machine capacity, which is cut, by action number, for example, No. 30 generator nodes cut off respectively capacity 50%,
60%, 70%, action number is respectively 1,2,3；No. 38 node generator nodes cut off capacity 50%, 60%, 70% respectively, move
It is respectively 25,26,27 to make number.In training process, action number is encoded using onehot coding modes, as Q-Learning
Training sample.Q-Learning builds network using the Tensorflow deep learning frames of Google publications, line number of going forward side by side
According to calculation processing.According to cutting machine control decision flow training pattern shown in Fig. 2 and obtain final result.
S103 determines control decision according to the operating status of the electric system, wherein the control decision is for protecting
Hinder the power system security stable operation.
After obtaining the operating status of electric system, it can select to adapt to Operation of Electric Systems shape from multiple control decisions
The control decision of state, so that electric system can be run to normal table.
Specifically, as shown in figure 3, cutting machine strategy generator deviation comparison diagram for 16-17 branch trouble differences.By Fig. 3
(b) it is the obtained control strategy of technical solution provided in an embodiment of the present invention, by technical solution provided in an embodiment of the present invention
After selected action excision generator, alternator speed deviation tends towards stability, and alternator speed deviation is suppressed in smaller model
Enclose fluctuation.Fig. 3 (a) random selection No. 30 generators of excision, alternator speed deviation is still continuously increased after excision, will finally be caused
Operation of power networks loss of stability.As it can be seen that control decision determined by the technical solution provided through the embodiment of the present invention, can make
It runs with obtaining electric system normal table.
Fig. 4 compare the different running method incision machine result of decision, wherein subgraph respectively represent 5-6,5-8,6-7,15-16,
The distribution of the machine result of decision is cut in the disturbance of 16-17,16-19,16-24 branch.Figure middle conductor represents return value, i.e. intensified learning model
Need the desired value learnt.Figure midpoint represents Q values result under different running method.In analytic process, rate of load condensate is randomly selected
The operation of power networks environmental information of 0.9-1.1, the small tactful effect of Q values is preferable, should pay the utmost attention to.Can be obtained by Fig. 4, Q Distribution values in
The both sides of return value line segment, and Q value overall distribution characteristics are consistent with return value line segment trend, while the corresponding Q of most of actions
Distribution value is more close.Fig. 4 (e)-(g) cuts the corresponding disturbance of machine strategy and is respectively positioned on No. 16 near nodals, Q Distribution value characteristic phases
Closely, return value line segment trend is similar, close for the different running method incision machine result of decision.Therefore, deeply learns mould
Type, which has, well adapts to ability.Since deeply learning framework contains multilayer neural network model, from data-driven angle
Neural network has certain generalization ability to data, is robustness from physical system angle.Based on described above, deeply
Study cuts machine Decision Control for power grid can adapt to power grid different running method to a certain extent, and acquired results have correctly
Property.
Compared with prior art, technical solution provided in an embodiment of the present invention can directly start with from electric network data, be based on
Trained deeply learning model obtains the learning value information in Operation of Electric Systems data in advance, then according to being obtained
The learning value information taken judges the operating status of electric system and provides reasonable control program；So as to a certain extent
Dependence of the electrical network analysis to physical model is avoided, and model support need not be readjusted for different running method different scenes
Structure, training effectiveness is high, with strong applicability；In addition, the embodiment of the present invention makes full use of data resource, it ensure that resource is not unrestrained
Take.
Second aspect, the embodiment of the present invention additionally provides a kind of device of processing electric power system data, as shown in figure 5, institute
Stating device includes：
Data obtaining module 510, for based on deeply learning model trained in advance, obtaining Operation of Electric Systems number
Learning value information in；
Operating status judgment module 520, the fortune for judging the electric system according to acquired learning value information
Row state；
Control decision determining module 530 determines control decision for the operating status according to the electric system, wherein
The control decision is for ensureing the power system security stable operation.
Optionally, the training process of the deeply learning model is as follows：
Deeply learning model is initialized, the deeply learning model is by deep learning model and intensified learning mould
Type is constituted；
Obtain electric network data；
Pretreatment is carried out to the electric network data obtained and forms sample data set；
It is concentrated from the sample data and chooses training data；
Utilize the feature of training data described in deep learning model extraction；
The feature extracted is analyzed using intensified learning model, obtains analysis result, and tie based on the analysis
Fruit judges whether to complete training；
If it is judged that complete training, the deeply learning model of training completion is obtained；
If it is judged that not complete training, the model ginseng of the deep learning model and intensified learning model is adjusted
Number returns to execute from the sample data and concentrates the step of choosing training data.
Optionally, the deep learning model is：Depth convolutional neural networks.
Optionally, the deep learning model is for improving initial data value density and increasing model generalization ability
Model.
Optionally, the intensified learning model is Q-Learning analysis models.
Compared with prior art, technical solution provided in an embodiment of the present invention can directly start with from electric network data, be based on
Trained deeply learning model obtains the learning value information in Operation of Electric Systems data in advance, then according to being obtained
The learning value information taken judges the operating status of electric system and provides reasonable control program；So as to a certain extent
Dependence of the electrical network analysis to physical model is avoided, and model support need not be readjusted for different running method different scenes
Structure, training effectiveness is high, with strong applicability；In addition, the embodiment of the present invention makes full use of data resource, it ensure that resource is not unrestrained
Take.
The third aspect, the embodiment of the present invention additionally provide a kind of electronic equipment, as shown in fig. 6, including processor 601, leading to
Believe interface 602, memory 603 and communication bus 604, wherein processor 601, communication interface 602, memory 603 pass through communication
Bus 604 completes mutual communication,
Memory 603, for storing computer program；
Processor 601 when for executing the program stored on memory 603, realizes the processing electricity described in first aspect
The method of Force system data.
The communication bus that above-mentioned electronic equipment is mentioned can be Peripheral Component Interconnect standard (Peripheral Component
Interconnect, PCI) bus or expanding the industrial standard structure (Extended Industry Standard
Architecture, EISA) bus etc..The communication bus can be divided into address bus, data/address bus, controlling bus etc..For just
It is only indicated with a thick line in expression, figure, it is not intended that an only bus or a type of bus.
Communication interface is for the communication between above-mentioned electronic equipment and other equipment.
Memory may include random access memory (Random Access Memory, RAM), can also include non-easy
The property lost memory (Non-Volatile Memory, NVM), for example, at least a magnetic disk storage.Optionally, memory may be used also
To be at least one storage device for being located remotely from aforementioned processor.
Above-mentioned processor can be general processor, including central processing unit (Central Processing Unit,
CPU), network processing unit (Network Processor, NP) etc.；It can also be digital signal processor (Digital Signal
Processing, DSP), it is application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), existing
It is field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic device, discrete
Door or transistor logic, discrete hardware components.
Compared with prior art, technical solution provided in an embodiment of the present invention can directly start with from electric network data, be based on
Trained deeply learning model obtains the learning value information in Operation of Electric Systems data in advance, then according to being obtained
The learning value information taken judges the operating status of electric system and provides reasonable control program；So as to a certain extent
Dependence of the electrical network analysis to physical model is avoided, and model support need not be readjusted for different running method different scenes
Structure, training effectiveness is high, with strong applicability；In addition, the embodiment of the present invention makes full use of data resource, it ensure that resource is not unrestrained
Take.
Fourth aspect additionally provides a kind of computer readable storage medium in another embodiment provided by the invention, should
Instruction is stored in computer readable storage medium, when run on a computer so that it is real that computer executes the above method
Apply the data search method described in example.
Compared with prior art, technical solution provided in an embodiment of the present invention can directly start with from electric network data, be based on
Trained deeply learning model obtains the learning value information in Operation of Electric Systems data in advance, then according to being obtained
The learning value information taken judges the operating status of electric system and provides reasonable control program；So as to a certain extent
Dependence of the electrical network analysis to physical model is avoided, and model support need not be readjusted for different running method different scenes
Structure, training effectiveness is high, with strong applicability；In addition, the embodiment of the present invention makes full use of data resource, it ensure that resource is not unrestrained
Take.
It should be noted that herein, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also include other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that
There is also other identical elements in process, method, article or equipment including the element.
Each embodiment in this specification is all made of relevant mode and describes, identical similar portion between each embodiment
Point just to refer each other, and each embodiment focuses on the differences from other embodiments.Especially for system reality
For applying example, since it is substantially similar to the method embodiment, so description is fairly simple, related place is referring to embodiment of the method
Part explanation.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the scope of the present invention.It is all
Any modification, equivalent replacement, improvement and so within the spirit and principles in the present invention, are all contained in protection scope of the present invention
It is interior.
Claims (10)
1. a kind of method of processing electric power system data, which is characterized in that the method includes：
Based on deeply learning model trained in advance, the learning value information in Operation of Electric Systems data is obtained；
The operating status of the electric system is judged according to acquired learning value information；
According to the operating status of the electric system, control decision is determined, wherein the control decision is for ensureing the electric power
System safe and stable operation.
2. according to the method described in claim 1, it is characterized in that, the training process of the deeply learning model is as follows：
Deeply learning model is initialized, the deeply learning model is by deep learning model and intensified learning model structure
At；
Obtain electric network data；
Pretreatment is carried out to the electric network data obtained and forms sample data set；
It is concentrated from the sample data and chooses training data；
Utilize the feature of training data described in deep learning model extraction；
The feature extracted is analyzed using intensified learning model, obtains analysis result, and sentence based on the analysis result
It is disconnected whether to complete to train；
If it is judged that complete training, the deeply learning model of training completion is obtained；
If it is judged that not complete training, the model parameter of the deep learning model and intensified learning model is adjusted, is returned
Receipt row concentrates the step of choosing training data from the sample data.
3. according to the method described in claim 2, it is characterized in that, the deep learning model is：Depth convolutional neural networks.
4. according to the method described in claim 2, it is characterized in that, the deep learning model is for improving initial data valence
It is worth density and increases the model of model generalization ability.
5. according to the method described in claim 4, it is characterized in that, the intensified learning model, which is Q-Learning, analyzes mould
Type.
6. a kind of device of processing electric power system data, which is characterized in that described device includes：
Data obtaining module, for based on deeply learning model trained in advance, obtaining in Operation of Electric Systems data
Learning value information；
Operating status judgment module, the operating status for judging the electric system according to acquired learning value information；
Control decision determining module determines control decision, wherein the control for the operating status according to the electric system
Decision processed is for ensureing the power system security stable operation.
7. device according to claim 5, which is characterized in that the training process of the deeply learning model is as follows：
Deeply learning model is initialized, the deeply learning model is by deep learning model and intensified learning model structure
At；
Obtain electric network data；
Pretreatment is carried out to the electric network data obtained and forms sample data set；
It is concentrated from the sample data and chooses training data；
Utilize the feature of training data described in deep learning model extraction；
The feature extracted is analyzed using intensified learning model, obtains analysis result, and sentence based on the analysis result
It is disconnected whether to complete to train；
If it is judged that complete training, the deeply learning model of training completion is obtained；
If it is judged that not complete training, the model parameter of the deep learning model and intensified learning model is adjusted, is returned
Receipt row concentrates the step of choosing training data from the sample data.
8. device according to claim 6, which is characterized in that the deep learning model is：Depth convolutional neural networks.
9. device according to claim 6, which is characterized in that the deep learning model is for improving initial data valence
It is worth density and increases the model of model generalization ability.
10. a kind of electronic equipment, which is characterized in that including processor, communication interface, memory and communication bus, wherein processing
Device, communication interface, memory complete mutual communication by communication bus；
Memory, for storing computer program；
Processor when for executing the program stored on memory, realizes any method and steps of claim 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810166848.XA CN108520472A (en) | 2018-02-28 | 2018-02-28 | A kind of method, apparatus and electronic equipment of processing electric power system data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810166848.XA CN108520472A (en) | 2018-02-28 | 2018-02-28 | A kind of method, apparatus and electronic equipment of processing electric power system data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108520472A true CN108520472A (en) | 2018-09-11 |
Family
ID=63433376
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810166848.XA Pending CN108520472A (en) | 2018-02-28 | 2018-02-28 | A kind of method, apparatus and electronic equipment of processing electric power system data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108520472A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109103936A (en) * | 2018-09-30 | 2018-12-28 | 南京铭越创信电气有限公司 | Optimal unit starting order calculation method after a kind of electric system is had a power failure on a large scale |
| CN109754075A (en) * | 2019-01-16 | 2019-05-14 | 中南民族大学 | Dispatching method, equipment, storage medium and the device of wireless sensor network node |
| CN109785289A (en) * | 2018-12-18 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of transmission line of electricity defect inspection method, system and electronic equipment |
| CN110516889A (en) * | 2019-09-03 | 2019-11-29 | 广东电网有限责任公司 | A kind of load Comprehensive Prediction Method and relevant device based on Q-learning |
| CN110535146A (en) * | 2019-08-27 | 2019-12-03 | 哈尔滨工业大学 | The Method for Reactive Power Optimization in Power of Policy-Gradient Reinforcement Learning is determined based on depth |
| CN111191529A (en) * | 2019-12-17 | 2020-05-22 | 中移（杭州）信息技术有限公司 | Method and system for processing abnormal work order |
| CN111327487A (en) * | 2018-12-14 | 2020-06-23 | 国网山西省电力公司信息通信分公司 | Power communication network running state monitoring method and device based on deep learning |
| CN111478292A (en) * | 2020-03-16 | 2020-07-31 | 中国电力科学研究院有限公司 | Fault removal method and system of power system based on deep reinforcement learning |
| CN111525587A (en) * | 2020-04-01 | 2020-08-11 | 中国电力科学研究院有限公司 | Reactive load situation-based power grid reactive voltage control method and system |
| CN111688104A (en) * | 2020-06-18 | 2020-09-22 | 诺兰特新材料(北京)有限公司 | Automatic burr removing system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105512808A (en) * | 2015-11-30 | 2016-04-20 | 武汉大学 | Power system transient stability assessment method based on big data |
| CN105676670A (en) * | 2014-11-18 | 2016-06-15 | 北京翼虎能源科技有限公司 | Method and system for processing energy data |
| CN107037278A (en) * | 2016-11-04 | 2017-08-11 | 国家电网公司 | A kind of substandard intelligent substation method for diagnosing faults of IEC61850 |
| CN107067190A (en) * | 2017-05-18 | 2017-08-18 | 厦门大学 | The micro-capacitance sensor power trade method learnt based on deeply |
-
2018
- 2018-02-28 CN CN201810166848.XA patent/CN108520472A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105676670A (en) * | 2014-11-18 | 2016-06-15 | 北京翼虎能源科技有限公司 | Method and system for processing energy data |
| CN105512808A (en) * | 2015-11-30 | 2016-04-20 | 武汉大学 | Power system transient stability assessment method based on big data |
| CN107037278A (en) * | 2016-11-04 | 2017-08-11 | 国家电网公司 | A kind of substandard intelligent substation method for diagnosing faults of IEC61850 |
| CN107067190A (en) * | 2017-05-18 | 2017-08-18 | 厦门大学 | The micro-capacitance sensor power trade method learnt based on deeply |
Non-Patent Citations (1)
| Title |
|---|
| 刘威等: "基于深度强化学习的电网紧急控制策略研究", 《中国电机工程学报》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109103936A (en) * | 2018-09-30 | 2018-12-28 | 南京铭越创信电气有限公司 | Optimal unit starting order calculation method after a kind of electric system is had a power failure on a large scale |
| CN111327487A (en) * | 2018-12-14 | 2020-06-23 | 国网山西省电力公司信息通信分公司 | Power communication network running state monitoring method and device based on deep learning |
| CN109785289A (en) * | 2018-12-18 | 2019-05-21 | 中国科学院深圳先进技术研究院 | A kind of transmission line of electricity defect inspection method, system and electronic equipment |
| CN109754075A (en) * | 2019-01-16 | 2019-05-14 | 中南民族大学 | Dispatching method, equipment, storage medium and the device of wireless sensor network node |
| CN110535146A (en) * | 2019-08-27 | 2019-12-03 | 哈尔滨工业大学 | The Method for Reactive Power Optimization in Power of Policy-Gradient Reinforcement Learning is determined based on depth |
| CN110516889A (en) * | 2019-09-03 | 2019-11-29 | 广东电网有限责任公司 | A kind of load Comprehensive Prediction Method and relevant device based on Q-learning |
| CN110516889B (en) * | 2019-09-03 | 2023-07-07 | 广东电网有限责任公司 | Load comprehensive prediction method based on Q-learning and related equipment |
| CN111191529A (en) * | 2019-12-17 | 2020-05-22 | 中移（杭州）信息技术有限公司 | Method and system for processing abnormal work order |
| CN111191529B (en) * | 2019-12-17 | 2023-04-28 | 中移（杭州）信息技术有限公司 | Method and system for processing abnormal worksheets |
| CN111478292A (en) * | 2020-03-16 | 2020-07-31 | 中国电力科学研究院有限公司 | Fault removal method and system of power system based on deep reinforcement learning |
| CN111478292B (en) * | 2020-03-16 | 2022-10-04 | 中国电力科学研究院有限公司 | Fault removal method and system of power system based on deep reinforcement learning |
| CN111525587A (en) * | 2020-04-01 | 2020-08-11 | 中国电力科学研究院有限公司 | Reactive load situation-based power grid reactive voltage control method and system |
| CN111525587B (en) * | 2020-04-01 | 2022-10-25 | 中国电力科学研究院有限公司 | Reactive load situation-based power grid reactive voltage control method and system |
| CN111688104A (en) * | 2020-06-18 | 2020-09-22 | 诺兰特新材料(北京)有限公司 | Automatic burr removing system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108520472A (en) | A kind of method, apparatus and electronic equipment of processing electric power system data | |
| CN110633745B (en) | Image classification training method and device based on artificial intelligence and storage medium | |
| EP2991004B1 (en) | Method and apparatus for labeling training samples | |
| CN107729465B (en) | Appraisal procedure, device and the electronic equipment of personage's danger level | |
| CN108009673A (en) | Novel load Forecasting Methodology and device based on deep learning | |
| CN112862092B (en) | Training method, device, equipment and medium for heterogeneous graph convolution network | |
| Branavan et al. | Non-linear monte-carlo search in civilization ii | |
| CN109102146B (en) | Electric power system risk assessment acceleration method based on multi-parameter linear programming | |
| CN111709406A (en) | Text line identification method and device, readable storage medium and electronic equipment | |
| Yu et al. | Forecasting house price index of China using dendritic neuron model | |
| CN112217207A (en) | Transient voltage stability margin prediction method and device | |
| CN107392311A (en) | The method and apparatus of sequence cutting | |
| CN106845991A (en) | A kind of rules process method and equipment | |
| CN111445065A (en) | Energy consumption optimization method and system for refrigeration group control of data center | |
| CN107506952A (en) | Appraisal procedure, device and the electronic equipment of hazard index | |
| Kuliev et al. | Monkey search algorithm for ECE components partitioning | |
| CN111598234B (en) | AI model training method, AI model using method, computer device, and storage medium | |
| CN110069772A (en) | Predict device, method and the storage medium of the scoring of question and answer content | |
| CN112402982A (en) | User cheating behavior detection method and system based on machine learning | |
| CN114726823B (en) | Domain name generation method, device and equipment based on generation countermeasure network | |
| CN116340777A (en) | Training method of log classification model, log classification method and device | |
| CN106897328A (en) | A kind of image search method and device | |
| CN116187310A (en) | Document-level relation extraction method, device, equipment and storage medium | |
| CN113747480B (en) | Processing method and device for 5G slice faults and computing equipment | |
| US8150785B2 (en) | Multi-modal green computing fusion using problem analytics |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180911 |
|
| WD01 | Invention patent application deemed withdrawn after publication |