CN112632846B - Power transmission section limit probability assessment method of power system and electronic equipment - Google Patents
Power transmission section limit probability assessment method of power system and electronic equipment Download PDFInfo
- Publication number
- CN112632846B CN112632846B CN202011266697.9A CN202011266697A CN112632846B CN 112632846 B CN112632846 B CN 112632846B CN 202011266697 A CN202011266697 A CN 202011266697A CN 112632846 B CN112632846 B CN 112632846B
- Authority
- CN
- China
- Prior art keywords
- layer
- neural network
- mode
- vector
- power
- 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.)
- Active
Links
- 230000005540 biological transmission Effects 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000003062 neural network model Methods 0.000 claims abstract description 51
- 238000011156 evaluation Methods 0.000 claims abstract description 36
- 239000013598 vector Substances 0.000 claims description 73
- 210000002569 neuron Anatomy 0.000 claims description 50
- 238000013528 artificial neural network Methods 0.000 claims description 45
- 238000010276 construction Methods 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 14
- 238000012549 training Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 239000000203 mixture Substances 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000012806 monitoring device Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 210000004205 output neuron Anatomy 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000013210 evaluation model Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 241000699670 Mus sp. Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
-
- 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
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00001—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by the display of information or by user interaction, e.g. supervisory control and data acquisition systems [SCADA] or graphical user interfaces [GUI]
-
- 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
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
-
- 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
-
- 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
- H02J3/04—Circuit arrangements for ac mains or ac distribution networks for connecting networks of the same frequency but supplied from different sources
- H02J3/06—Controlling transfer of power between connected networks; Controlling sharing of load between connected networks
-
- 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
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- 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
Abstract
The invention discloses a power transmission section limit probability assessment method of a power system and electronic equipment, wherein the method collects current running state data of the power system, obtains input characteristic data according to the current running state data, and inputs the input characteristic data into a pre-constructed depth uncertainty neural network model xi d In (a) and (b); based on depth uncertainty neural network model ζ d And outputting the predicted TTC probability distribution of the given transmission section of the power system. The method can output the probability distribution of the TTC of the power transmission section of the power system, and provides an implementation method for power transmission section limit probability evaluation under the consideration of uncertainty.
Description
Technical Field
The invention belongs to the technical field of safety analysis of power systems, and particularly relates to a power transmission section limit probability evaluation method of a power system based on a deep uncertainty neural network.
Background
The power system is one of the largest energy systems. In a traditional manner, system monitoring, security analysis and system control are the three primary steps of a dispatcher in maintaining the security and stability of a power system. Through system monitoring, a dispatcher can acquire the latest information of the power grid in real time and directly judge whether the current power grid is in a normal running state. And then, through safety analysis, estimating the current operation condition of the power grid, and detecting weak links possibly occurring in the potential operation scene of the power grid based on the current operation scene, wherein the N-1 fault condition is mainly considered. Finally, when the power system is judged to be in an abnormal condition or weak links appear in a potential operation scene, a dispatcher needs to perform corresponding dispatching control to avoid the safety problem of the power system. The complex power system has a huge amount of running states, a scheduler cannot conduct overall monitoring, and a common method for personnel in a power system scheduling center to conduct operation mode programming, stable quota formulation and monitoring of a scheduling table is to pay attention to a series of power transmission sections, and focus calculation and monitoring are conducted on the sections which are easier to overrun. In the power grid safety evaluation, the evaluation of the limit transmission capacity (Total Transfer Capability, TTC) of a power transmission section is one of the most central links.
The data-driven deep learning model shows higher robustness, faster calculation speed and strong nonlinear fitting capability in the TTC evaluation problem of the power transmission section, and is increasingly applied in the TTC evaluation problem in recent years. However, most of existing TTC evaluation models based on deep learning only output a single predicted value, lack probability information, and consider high uncertainty caused by continuous access of new energy and flexible load, so that a dispatcher is difficult to make a risk decision based on the single TTC predicted value.
Disclosure of Invention
The invention aims to solve the technical problem of providing a power transmission section limit probability evaluation method of a power system based on a deep uncertainty neural network, and solves the problems that the existing TTC evaluation model based on deep learning only outputs a single predicted value, lacks probability information and is difficult to carry out risk decision based on the single TTC predicted value.
In order to solve the technical problems, the invention adopts the following technical scheme:
the power transmission section limit probability evaluation method of the power system based on the neural network comprises the steps of collecting current running state data of the power system, obtaining input characteristic data according to the current running state data, and inputting the input characteristic data into a pre-constructed depth uncertainty neural network model xi d In (a) and (b); based on depth uncertainty neural network model ζ d Outputting predicted TTC probability distribution of a given transmission section of the power system;
the obtaining of the input characteristic data comprises the following substeps:
(1-1) collecting power system characteristic data from a power system monitoring device to form an input characteristic x= [ x ] of a depth uncertainty neural network model for power system transmission section limit probability evaluation q ,x p], wherein ,xq Representing the topological structure feature vector of the power grid, x p The method is characterized in that the characteristic vector of the power grid tide state is represented, x is an input characteristic vector, and the construction mode is as follows:
(1-1-1) Power grid topology feature vector x q The specific composition mode of (a) is as follows:
wherein ,represents the admittance of the s-th line;
(1-1-2) Power grid tidal current State feature vector x p The specific composition mode of (a) is as follows:
wherein ,the characteristic quantity of the power grid tide state is the f-th power grid tide state characteristic quantity in the power grid running state, and the power grid tide state characteristic quantity consists of the active power of the generator, the generator terminal voltage amplitude and the active power of the load node;
(1-2) pair x p The method comprises the following specific steps of:
(1-2-1) for the f-th grid tidal current state feature quantity, calculating the average absolute error for all the samples
Wherein N is the number of samples of the running state of the power grid,is the characteristic quantity, m, of the power flow state of the f power grid in the running state of the i power grid f The average value of the characteristic quantity of the f power flow state of the power grid;
(1-2-2) pairStandardized as ∈>
(1-2-3) obtaining a normalized power grid power flow state vector for the ith power grid running state based on the sample average absolute error calculation in (1-2-1) and the normalization method in (1-2-2)
(1-2-4) combining the grid power flow state characteristic vector in the step (1-1-1) with the standardized grid power flow state vector in the step (1-2-3) to form standardized input characteristic data x' = (x) q ,z p )。
Preferably, the depth uncertainty neural network model ζ d The construction method of the method comprises the following steps:
(2-1) construction of deep neural network model ζ d The input layer of (2) the feature data x' = (x) after the preprocessing in step (1-2) q ,z p ) As a model of a deep uncertainty neural network ζ d Is a layer of the input layer;
(2-2) construction of a depth uncertainty neural network model ζ d Is a hidden layer of (2);
(2-3) construction of depth uncertainty neural network ζ d Is a pattern layer of (1);
(2-4) construction of depth uncertainty neural network ζ d Is a summation layer of (2);
(2-5) construction of depth uncertainty neural network ζ d Is a regularized layer of (a);
(2-6) construction of depth uncertainty neural network ζ d An output layer of (2);
(3) Training depth uncertainty neural network ζ d Obtaining a depth uncertainty neural network model xi d 。
Preferably, the mode layers include 1 mode vector layer and 1 mode outputThe layer, the mode vector layer is connected with the last hidden layer h s Thereafter, the mode vector is defined by n of the mode layer hs The neuron and the last hidden layer h s The connection weight parameters between the two modes do not share any connection weight parameter, the mode vectors are not shared between the modes, and the last hidden layer h is assumed s N is the number of neurons of (2) hs The number of neurons of the pattern vector layer is n hs X p x v, where p is the number of modes, v is the number of mode vectors in each mode, and the length of the mode vector is n hs ；
The mode output layer is connected behind the mode vector layer, the number of neurons of the mode output layer is p, and each mode vector corresponds to one mode output neuron; the vector output by the last hidden layer is recorded as f, and the kth mode vector of the ith mode is recorded asMode output corresponding to the ith mode and the kth mode vector +.>From f and->Gaussian kernel function calculations between them to describe the similarity between the two:
where σ is the smooth index.
Preferably, the depth uncertainty neural network model ζ d The summation layer of the model is 1 layer, the number of nodes is the number p of modes, and each neuron is used for calculating the weighted average of the output values of the mode output layer neurons corresponding to the similar modes in the mode layer.
Preferably, the depth uncertainty neural network model ζ d The regularization layer of the (C) is 1 layer, the number of nodes is the number p of modes, and the regularization layer isThe nodes are in one-to-one correspondence with the nodes of the summation layer and are used for regularizing the output of the summation layer to form the posterior probability density.
Preferably, the depth uncertainty neural network model ζ d The final output of (a) is the probability that the predicted TTC value falls in different TTC intervals, and a depth uncertainty neural network model xi is set d Is y, then it is of the form:
y＝[p(TTC∈[I 11 ,I 12 ]|x′),...,p(TTC∈[I i1 ,I i2 ]|x′),...,p(TTC∈[I p1 ,I p2 ]|x′)]。
preferably, the training method of the parameters in the depth uncertainty neural network model comprises the following steps:
(3-1) the parameter set to be trained is B, comprising:
a. connection weight w between input layer neuron and first hidden layer neuron 1 Input layer neuron to first hidden layer neuron bias vector b 1 ；
b. Connection weight w between the j-th layer neuron and the j+1-th layer neuron from the input layer among the s hidden layers j Bias vector b for layer j neurons to layer j+1 neurons j ，j≥2；
c. The last hidden layer neuron, i.e. the connection weight w between the s+1st layer and the mode layer neuron s+1 ；
(3-2) setting a loss function as follows:
wherein, for the kth sample, x (k) Is input to the depth uncertainty neural network ζ d Is p (TTC. Epsilon.I) i |x (k) ) Is the true value of the i-th mode, and takes the value of only 0 or 1, h ∑Pi (TTC∈I i |x (k) ) Is an estimate of the class i mode summation layer, gamma|B| 2 Is a regular term connecting the weight matrix;
(3-3) utilization of deep nervesTraining method of network model, for depth uncertainty neural network xi of step (2) d Training based on the loss function in (3-2) to obtain a depth uncertainty neural network model xi d A parameter set B of the network;
(3-4) obtaining a deep uncertainty neural network model ζ according to the network parameters in the step (3-3) d 。
The invention also provides electronic equipment, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the power transmission section limit probability evaluation method of the power system based on the neural network when executing the computer program.
The invention also provides a medium which stores a computer program, and the computer program can realize the power transmission section limit probability evaluation method of the power system based on the neural network when being executed by a processor.
The invention provides a power transmission section limit probability evaluation method and device of a power system based on a depth uncertainty neural network, which take the problem that a dispatcher is difficult to carry out risk decision based on a single TTC predicted value of a power transmission section into consideration, and has the following beneficial effects:
based on the big data of the power system, a data-driven deep uncertainty neural network model is established, probability distribution of a TTC (time to failure) of a power transmission section of the power system can be output, and an implementation method is provided for power transmission section limit probability evaluation under the consideration of uncertainty.
The method can be applied to the field of safety evaluation of the power system, and the accuracy and the robustness of the safety evaluation are improved.
The method provides richer prediction information and decision basis for power system operation scheduling personnel, and improves the safety of power system scheduling.
The specific technical scheme and the beneficial effects of the invention will be described in detail in the following specific embodiments.
Drawings
The invention is further described with reference to the drawings and detailed description which follow:
fig. 1 is a flowchart of a power transmission section limit probability evaluation method of a power system.
Detailed Description
The technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments. The following description of at least one exemplary embodiment is merely exemplary in nature and is in no way intended to limit the invention, its application, or uses. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
The power transmission section limit probability evaluation method based on the depth uncertainty neural network provided by the embodiment comprises the following steps:
(1) And determining input characteristic data x of a safety early warning model of the power system and limit transmission capacity TTC of a given transmission section, wherein the transmission section is manually given by operation mode personnel.
The method comprises the following specific steps:
(1-1) collecting power system characteristic data from a power system monitoring device to form an input characteristic x= [ x ] of a deep neural network model for power system safety pre-warning q ,x p], wherein ,xq Representing the topological structure feature vector of the power grid, x p The characteristic vector of the power grid tide state is represented, x is an input characteristic vector, and the specific construction mode is as follows:
(1-1-1) Power grid topology feature vector x q The specific composition mode of (a) is as follows:
wherein ,represents the admittance of the s-th line;
(1-1-2) Power grid tidal current State feature vector x p The specific composition mode of (a) is as follows:
wherein ,the characteristic quantity of the power grid tide state is the f-th power grid tide state characteristic quantity in the power grid running state, and the power grid tide state characteristic quantity consists of the active power of the generator, the generator terminal voltage amplitude and the active power of the load node;
(1-2) to avoid reliance on metric units, x is required p The method comprises the following specific steps of:
(1-2-1) for the f-th grid tidal current state feature quantity, calculating the average absolute error for all the samples
Wherein N is the number of samples of the running state of the power grid,is the characteristic quantity, m, of the power flow state of the f power grid in the running state of the i power grid f The average value of the characteristic quantity of the f power flow state of the power grid;
(1-2-2) pairStandardized as ∈>
(1-2-3) obtaining a normalized power grid power flow state vector for the ith power grid running state based on the sample average absolute error calculation in (1-2-1) and the normalization method in (1-2-2)
(1-2-4) combining the grid power flow state characteristic vector in the step (1-1-1) with the standardized grid power flow state vector in the step (1-2-3) to form a standardized input characteristic vector x' = (x) q ,z p )；
(1-3) calculating to obtain a section TTC value P corresponding to the ith power grid running state by using a simulation tool TTCi As the label corresponding to the input characteristic of the ith power grid running state;
(1-4) forming a sample (P) corresponding to the ith grid operating condition TTCi ；x′ (i) )；
(2) Constructing a deep uncertainty neural network model xi for power system transmission section limit probability evaluation d The method comprises the following specific steps of:
(2-1) construction of deep neural network model ζ d Is a layer of the input layer:
and (3) preprocessing the characteristic data x' = (x) in the step (1-2) q ,z p ) Input depth uncertainty neural network model ζ d Becomes a depth uncertainty neural network model xi d The input layer is used for receiving data input;
(2-2) construction of a depth uncertainty neural network model ζ d Is a hidden layer of s:
depth uncertainty neural network model ζ d Comprising s hidden layers which, in this patent,s takes value as 4, and data obtained from the input layer are subjected to layer-by-layer data reduction and high-order feature extraction by utilizing s hidden layers, wherein the 1 st hidden layer h 1 Is the input layer x' of the whole model, the w-th hidden layer h w The input of (2) is h w-1 The output of the layers, w=2, …, s, the number of hidden layers s and the number of neurons of each hidden layer are set according to the prediction precision requirement, and the number of neurons of each layer is gradually decreased;
(2-3) construction of depth uncertainty neural network ζ d The pattern layer of (2) is divided into the following two steps:
(2-3-1) construction of deep uncertainty neural network ζ d Mode vector layer among the mode layers:
depth uncertainty neural network model ζ d The pattern layer of (2) contains 1 pattern vector layer, and the pattern vector layer is connected with the last hidden layer h s After that, the process is performed. The meaning of a pattern is a section of TTC, e.g. pattern I is a section of TTC [ I ] i1 ,I i2 ]Is set by human. The mode vector is defined by n of the mode layer hs The neuron and the last hidden layer h s The connection weight parameters between the two modes are composed, meaning that one of the feature vectors corresponding to each mode is needed to be obtained through training, any connection weight parameter is not shared between every two mode vectors, and the mode vectors are not shared between the modes. If the last hidden layer h s N is the number of neurons of (2) hs The number of neurons of the pattern vector layer is n hs X p x v, wherein p is the number of modes contained in the mode layer, the more the number of modes is, the more continuous the TTC probability distribution finally output by the neural network is, the less the number of modes is, the more discrete the TTC probability distribution finally output by the neural network is, the number of modes is 24 in the patent, v is the number of mode vectors in each mode, the value is 10 in the patent, and the length of the mode vectors is n hs The value in this patent is 12.
(2-3-2) construction of deep uncertainty neural network ζ d Mode output layer in the mode layer:
depth uncertainty neural network model ζ d The pattern layer of the pattern layer comprises 1 pattern output layer, the pattern output layer is connected behind the pattern vector layer, the number of neurons of the pattern output layer is p, and each pattern vector corresponds to one pattern output neuron; the vector output by the last hidden layer is recorded as f, and the kth mode vector of the ith mode is recorded asMode output corresponding to the ith mode and the kth mode vector +.>From f and->Gaussian kernel function calculations between them to describe the similarity between the two:
wherein sigma is a smooth index, designated by human, and is a value of 0.5 in this patent.
(2-4) construction of depth uncertainty neural network ζ d Is provided for the sum layer of:
depth uncertainty neural network model ζ d Is 1 layer, the number of nodes is the number of modes p, each neuron is used for calculating the weighted average of the output values of the mode output layer neurons corresponding to the same type of modes in the mode layer, for example, the output h of the ith node in the summation layer ∑Pi Is a weighted average of all output values of the i-th type mode in the mode layer:
wherein the weight ω k The calculation is performed by the following way:
(2-5) construction of depth uncertainty neural network ζ d Is used for the regularization layer:
depth uncertainty neural network model ζ d The regularization layer of the model is 1 layer, the number of the nodes is the number p of the modes, and the nodes of the regularization layer are in one-to-one correspondence with the nodes of the summation layer and are used for regularizing the output of the summation layer to form posterior probability density. For example, the output of the ith node in the regularization layer is the posterior probability density corresponding to the ith mode, or the cross section TTC value corresponding to the input feature vector falls in the ith mode value interval [ I ] i1 ,I i2 ]Posterior probability in:
(2-6) construction of depth uncertainty neural network ζ d Is provided with an output layer of:
depth uncertainty neural network model ζ d The output layers of the regularized layer are combined together to form the depth uncertainty neural network model xi d Representing the discrete probability distribution of TTC values corresponding to given sections of the current input power grid running state, specifically the probability of the predicted TTC values falling in different TTC intervals, and setting a depth uncertainty neural network model xi d Is y, then it is of the form:
y＝[p(TTC∈[I 11 ,I 12 ]|x′),...,p(TTC∈[I i1 ,I i2 ]|x′),...,p(TTC∈[I p1 ,I p2 ]|x′)]
(3) Training depth uncertainty neural network ζ d The specific steps are as follows:
(3-1) the parameter set to be trained is B, which specifically comprises:
a. connection weight w between input layer neuron and first hidden layer neuron 1 Input layer neuron to first hidden layer neuron bias vector b 1 ；
b. s number ofConnection weight w between a jth layer neuron and a j+1th layer neuron from an input layer in the hidden layer j Bias vector b for layer j neurons to layer j+1 neurons j ，j≥2；
c. The last hidden layer neuron, i.e. the connection weight w between the s+1st layer and the mode layer neuron s+1
(3-2) setting a loss function as follows:
wherein, for the kth sample, x (k) Is input to the depth uncertainty neural network ζ d Is p (TTC. Epsilon.I) i |x (k) ) Is the true value of the i-th mode, and takes the value of only 0 or 1, h ∑Pi (TTC∈I i |x (k) ) Is an estimate of the class i mode summation layer, gamma|B| 2 Is a regular term for connecting weight matrixes, wherein the parameter gamma is set by people, and the value of the parameter gamma is 0.0001 in the patent;
(3-3) training the depth uncertainty neural network xi of the step (2) by using a depth neural network model training method d Training based on the loss function in (3-2) to obtain a depth uncertainty neural network model xi d A parameter set B of the network;
(3-4) obtaining a deep uncertainty neural network model ζ according to the network parameters in the step (3-3) d ；
(4) Collecting the current running state of the power system from a power system monitoring device, forming a characteristic vector in the mode of (1-1), preprocessing the characteristic vector of the power system by using the data preprocessing method in the step (1-2) to obtain characteristic data for inputting a network, and inputting the characteristic data into the depth uncertainty neural network model xi in the step (3) d In the output power system, the predicted TTC probability distribution of a given transmission section
The power transmission section limit probability evaluation method based on the deep uncertainty neural network has the advantages that the deep neural network model is adopted in the method, and an implementation method is provided for power transmission section limit probability evaluation under the consideration of uncertainty. The method disclosed by the invention considers the problem that a dispatcher is difficult to carry out risk decision based on a single TTC predicted value of the power transmission section, introduces a deep learning method considering uncertainty to carry out probability evaluation on the TTC of the power transmission section, overcomes the defect, and provides more abundant predicted information and decision basis for the dispatcher. The method can be applied to the field of safety evaluation of the power system, improves the accuracy and the robustness of the safety evaluation, provides scheduling basis for power system operation scheduling personnel, and improves the safety of power system scheduling.
Example two
An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the depth uncertainty neural network-based power system transmission section limit probability assessment method of embodiment one when executing the computer program.
The electronic device in the embodiment of the present invention may include, but is not limited to, a mobile terminal such as a mobile phone, a notebook computer, a PDA (personal digital assistant), a PAD (tablet computer), etc., and a fixed terminal such as a desktop computer, etc.
The electronic device may include a processing means (e.g., a central processing unit) that may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) or a program loaded from a storage means into a Random Access Memory (RAM). In the RAM, various programs and data required for the operation of the electronic device are also stored. The processing device, ROM and RAM are connected to each other via a bus. An input/output (I/O) interface is also connected to the bus.
In general, the following systems may be connected to the I/O interface: input devices including, for example, touch screens, touch pads, keyboards, mice, etc.; output devices including, for example, liquid Crystal Displays (LCDs), speakers, vibrators, etc.; storage devices including, for example, magnetic tape, hard disk, etc.; a communication device. The communication means may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data.
A computer program, carried on a computer readable medium, contains program code for performing the method. In such an embodiment, the computer program may be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from ROM. The above-described functions defined in the methods of the embodiments of the present disclosure are performed when the computer program is executed by a processing device.
It should be noted that the computer readable medium described in the present disclosure may be a computer readable signal medium or a computer readable medium, or any combination of the two. The computer readable medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor apparatus, system, or device, or a combination of any of the foregoing. More specific examples of the computer-readable medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution apparatus, system, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution apparatus, system, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.
The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.
The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring at least two internet protocol addresses; sending a node evaluation request comprising the at least two internet protocol addresses to node evaluation equipment, wherein the node evaluation equipment selects an internet protocol address from the at least two internet protocol addresses and returns the internet protocol address; receiving an Internet protocol address returned by the node evaluation equipment; wherein the acquired internet protocol address indicates an edge node in the content distribution network.
Alternatively, the computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: receiving a node evaluation request comprising at least two internet protocol addresses; selecting an internet protocol address from the at least two internet protocol addresses; returning the selected internet protocol address; wherein the received internet protocol address indicates an edge node in the content distribution network.
Computer program code for carrying out operations of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, step Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).
While the invention has been described in terms of specific embodiments, it will be appreciated by those skilled in the art that the invention is not limited to the specific embodiments described above. Any modifications which do not depart from the functional and structural principles of the present invention are intended to be included within the scope of the appended claims.
Claims (9)
1. The power transmission section limit probability evaluation method of the power system based on the neural network is characterized by comprising the following steps of: collecting current running state data of an electric power system, obtaining input characteristic data according to the current running state data, and inputting the input characteristic data into a pre-constructed depth uncertainty neural network model xi d In (a) and (b); based on depth uncertainty neural network model ζ d Outputting predicted TTC probability distribution of a given transmission section of the power system;
the obtaining of the input characteristic data comprises the following substeps:
(1-1) collecting power system characteristic data from a power system monitoring device to form an input characteristic x= [ x ] of a depth uncertainty neural network model for power system transmission section limit probability evaluation q ,x p], wherein ,xq Representing the topological structure feature vector of the power grid, x p The method is characterized in that the characteristic vector of the power grid tide state is represented, x is an input characteristic vector, and the construction mode is as follows:
(1-1-1) Power grid topology feature vector x q The specific composition mode of (a) is as follows:
wherein ,represents the admittance of the s-th line;
(1-1-2) Power grid tidal CurrentState feature vector x p The specific composition mode of (a) is as follows:
wherein ,the characteristic quantity of the power grid tide state is the f-th power grid tide state characteristic quantity in the power grid running state, and the power grid tide state characteristic quantity consists of the active power of the generator, the generator terminal voltage amplitude and the active power of the load node;
(1-2) pair x p The method comprises the following specific steps of:
(1-2-1) for the f-th grid tidal current state feature quantity, calculating the average absolute error for all the samples
Wherein N is the number of samples of the running state of the power grid,is the characteristic quantity, m, of the power flow state of the f power grid in the running state of the i power grid f The average value of the characteristic quantity of the f power flow state of the power grid;
(1-2-2) pairStandardized as ∈>
(1-2-3) obtaining a normalized power grid power flow state vector for the ith power grid running state based on the sample average absolute error calculation in (1-2-1) and the normalization method in (1-2-2)
(1-2-4) combining the grid power flow state characteristic vector in the step (1-1-1) with the standardized grid power flow state vector in the step (1-2-3) to form standardized input characteristic data x' = (x) q ,z p )。
2. The power transmission section limit probability evaluation method based on the neural network according to claim 1, wherein: the depth uncertainty neural network model ζ d The construction method of the method comprises the following steps:
(2-1) construction of deep neural network model ζ d The input layer of (2) the feature data x' = (x) after the preprocessing in step (1-2) q ,z p ) As a model of a deep uncertainty neural network ζ d Is a layer of the input layer;
(2-2) construction of a depth uncertainty neural network model ζ d Is a hidden layer of (2);
(2-3) construction of depth uncertainty neural network ζ d Is a pattern layer of (1);
(2-4) construction of depth uncertainty neural network ζ d Is a summation layer of (2);
(2-5) construction of depth uncertainty neural network ζ d Is a regularized layer of (a);
(2-6) construction of depth uncertainty neural network ζ d An output layer of (2);
(3) Training depth uncertainty neural network ζ d Is a deep oneUncertainty neural network model ζ d 。
3. The power transmission section limit probability evaluation method based on the neural network according to claim 2, characterized in that: the mode layers comprise 1 mode vector layer and 1 mode output layer, and the mode vector layer is connected with the last hidden layer h s Thereafter, the mode vector is defined by n of the mode layer hs The neuron and the last hidden layer h s The connection weight parameters between the two modes do not share any connection weight parameter, the mode vectors are not shared between the modes, and the last hidden layer h is assumed s N is the number of neurons of (2) hs The number of neurons of the pattern vector layer is n hs X p x v, where p is the number of modes, v is the number of mode vectors in each mode, and the length of the mode vector is n hs ；
The mode output layer is connected behind the mode vector layer, the number of neurons of the mode output layer is p, and each mode vector corresponds to one mode output neuron; the vector output by the last hidden layer is recorded as f, and the kth mode vector of the ith mode is recorded asMode output corresponding to the ith mode and the kth mode vector +.>From f and->Gaussian kernel function calculations between them to describe the similarity between the two:
where σ is the smooth index.
4. The power transmission section limit probability evaluation method based on the neural network according to claim 2, characterized in that: depth uncertainty neural network model ζ d The summation layer of the model is 1 layer, the number of nodes is the number p of modes, and each neuron is used for calculating the weighted average of the output values of the mode output layer neurons corresponding to the similar modes in the mode layer.
5. The power transmission section limit probability evaluation method based on the neural network according to claim 4, wherein: depth uncertainty neural network model ζ d The regularization layer of the model is 1 layer, the number of the nodes is the number p of the modes, and the nodes of the regularization layer are in one-to-one correspondence with the nodes of the summation layer and are used for regularizing the output of the summation layer to form posterior probability density.
6. The power transmission section limit probability evaluation method based on the neural network according to claim 2, characterized in that: depth uncertainty neural network model ζ d The final output of (a) is the probability that the predicted TTC value falls in different TTC intervals, and a depth uncertainty neural network model xi is set d Is y, then it is of the form:
y＝[p(TTC∈[I 11 ,I 12 ]|x′),...,p(TTC∈[I i1 ,I i2 ]|x′),...,p(TTC∈[I p1 ,I p2 ]|x′)]。
7. the power transmission section limit probability evaluation method based on the neural network according to claim 2, characterized in that: the training method of the parameters in the depth uncertainty neural network model comprises the following steps:
(3-1) the parameter set to be trained is B, comprising:
a. connection weight w between input layer neuron and first hidden layer neuron 1 Input layer neuron to first hidden layer neuron bias vector b 1 ；
b. The j from the input layer in the s hidden layersConnection weight w between layer neuron and j+1th layer neuron j Bias vector b for layer j neurons to layer j+1 neurons j ，j≥2；
c. The last hidden layer neuron, i.e. the connection weight w between the s+1st layer and the mode layer neuron s+1 ；
(3-2) setting a loss function as follows:
wherein, for the kth sample, x (k) Is input to the depth uncertainty neural network ζ d Is p (TTC. Epsilon.I) i |x (k) ) Is the true value of the i-th mode, and takes the value of only 0 or 1, h ∑Pi (TTC∈I i |x (k) ) Is an estimate of the class i mode summation layer, gamma|B| 2 Is a regular term connecting the weight matrix;
(3-3) training the depth uncertainty neural network xi of the step (2) by using a depth neural network model training method d Training based on the loss function in (3-2) to obtain a depth uncertainty neural network model xi d A parameter set B of the network;
(3-4) obtaining a deep uncertainty neural network model ζ according to the network parameters in the step (3-3) d 。
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized by: the processor, when executing the computer program, implements the steps of the power transmission section limit probability assessment method for a power system based on a neural network as claimed in any one of claims 1 to 7.
9. A storage medium storing a computer program, wherein the computer program, when executed by a processor, is capable of implementing the neural network-based power system transmission section limit probability assessment method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011266697.9A CN112632846B (en) | 2020-11-13 | 2020-11-13 | Power transmission section limit probability assessment method of power system and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011266697.9A CN112632846B (en) | 2020-11-13 | 2020-11-13 | Power transmission section limit probability assessment method of power system and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112632846A CN112632846A (en) | 2021-04-09 |
| CN112632846B true CN112632846B (en) | 2023-10-24 |
Family
ID=75303278
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011266697.9A Active CN112632846B (en) | 2020-11-13 | 2020-11-13 | Power transmission section limit probability assessment method of power system and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112632846B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103246806A (en) * | 2013-04-25 | 2013-08-14 | 浙江大学 | Operation risk evaluation method comprising wind- power plant electric system |
| CN109117951A (en) * | 2018-01-15 | 2019-01-01 | 重庆大学 | Probabilistic Load Flow on-line calculation method based on BP neural network |
| CN109638838A (en) * | 2019-01-21 | 2019-04-16 | 广东电网有限责任公司 | The recognition methods of power grid key sections, device and electronic equipment |
| CN110110434A (en) * | 2019-05-05 | 2019-08-09 | 重庆大学 | A kind of initial method that Probabilistic Load Flow deep neural network calculates |
| CN110991741A (en) * | 2019-12-02 | 2020-04-10 | 中国南方电网有限责任公司 | Section constraint probability early warning method and system based on deep learning |
| CN112001066A (en) * | 2020-07-30 | 2020-11-27 | 四川大学 | Deep learning-based method for calculating limit transmission capacity |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101527110B1 (en) * | 2009-04-13 | 2015-06-16 | 삼성전자주식회사 | Apparatus and method for power control in distributed multiple input multiple output wireless communication system |
| US11556794B2 (en) * | 2017-08-31 | 2023-01-17 | International Business Machines Corporation | Facilitating neural networks |
| SG11202105629SA (en) * | 2018-12-04 | 2021-06-29 | Google Llc | Generating integrated circuit floorplans using neural networks |
-
2020
- 2020-11-13 CN CN202011266697.9A patent/CN112632846B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103246806A (en) * | 2013-04-25 | 2013-08-14 | 浙江大学 | Operation risk evaluation method comprising wind- power plant electric system |
| CN109117951A (en) * | 2018-01-15 | 2019-01-01 | 重庆大学 | Probabilistic Load Flow on-line calculation method based on BP neural network |
| CN109638838A (en) * | 2019-01-21 | 2019-04-16 | 广东电网有限责任公司 | The recognition methods of power grid key sections, device and electronic equipment |
| CN110110434A (en) * | 2019-05-05 | 2019-08-09 | 重庆大学 | A kind of initial method that Probabilistic Load Flow deep neural network calculates |
| CN110991741A (en) * | 2019-12-02 | 2020-04-10 | 中国南方电网有限责任公司 | Section constraint probability early warning method and system based on deep learning |
| CN112001066A (en) * | 2020-07-30 | 2020-11-27 | 四川大学 | Deep learning-based method for calculating limit transmission capacity |
Non-Patent Citations (2)
| Title |
|---|
| N-1静态安全潮流约束下的输电断面有功潮流控制;李响等;电网技术;第29卷(第3期);第29-32页 * |
| 含大型太阳能发电系统的极限传输容量概率计算;王敏等;电力系统自动化;第34卷(第7期);第31-35页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112632846A (en) | 2021-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022077587A1 (en) | Data prediction method and apparatus, and terminal device | |
| CN112348660B (en) | Method and device for generating risk warning information and electronic equipment | |
| CN112529727A (en) | Micro-grid energy storage scheduling method, device and equipment based on deep reinforcement learning | |
| CN112541124B (en) | Method, apparatus, device, medium and program product for generating a multitasking model | |
| CN111611085B (en) | Yun Bian collaboration-based man-machine hybrid enhanced intelligent system, method and device | |
| CN111523640A (en) | Training method and device of neural network model | |
| CN114490065A (en) | Load prediction method, device and equipment | |
| CN112766402A (en) | Algorithm selection method and device and electronic equipment | |
| CN116684330A (en) | Traffic prediction method, device, equipment and storage medium based on artificial intelligence | |
| CN114118570A (en) | Service data prediction method and device, electronic equipment and storage medium | |
| WO2021077977A1 (en) | Method for analyzing transaction data in wind power bidding market, device, apparatus, and medium | |
| CN113825165B (en) | 5G slice network congestion early warning method and device based on time diagram network | |
| CN113886454A (en) | Cloud resource prediction method based on LSTM-RBF | |
| CN112632846B (en) | Power transmission section limit probability assessment method of power system and electronic equipment | |
| CN115952928A (en) | Short-term power load prediction method, device, equipment and storage medium | |
| CN115275975A (en) | Method and device for determining electric power data matching degree of optical storage charging station | |
| CN115481767A (en) | Operation data processing method and device for power distribution network maintenance and computer equipment | |
| CN114970357A (en) | Energy-saving effect evaluation method, system, device and storage medium | |
| CN115662510A (en) | Method, device and equipment for determining causal parameters and storage medium | |
| CN114971053A (en) | Training method and device for online prediction model of network line loss rate of low-voltage transformer area | |
| CN110889635B (en) | Method for performing emergency drilling on food safety event processing | |
| CN113570204A (en) | User behavior prediction method, system and computer equipment | |
| CN116681185B (en) | Load prediction method, device and chip equipment | |
| CN111709583B (en) | User retention time generation method, device, electronic equipment and medium | |
| Yang | ANN application techniques for power system stability estimation |
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 |