CN117439068A - Voltage sag estimation method, medium and system in large-scale power grid - Google Patents
Voltage sag estimation method, medium and system in large-scale power grid Download PDFInfo
- Publication number
- CN117439068A CN117439068A CN202311399848.1A CN202311399848A CN117439068A CN 117439068 A CN117439068 A CN 117439068A CN 202311399848 A CN202311399848 A CN 202311399848A CN 117439068 A CN117439068 A CN 117439068A
- Authority
- CN
- China
- Prior art keywords
- power grid
- data
- model
- voltage sag
- voltage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 17
- 230000009467 reduction Effects 0.000 claims abstract description 10
- 238000005291 chaos (dynamical) Methods 0.000 claims abstract description 4
- 238000004088 simulation Methods 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 11
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000005457 optimization Methods 0.000 claims description 9
- 230000009466 transformation Effects 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000000638 solvent extraction Methods 0.000 claims description 4
- 230000001052 transient effect Effects 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 3
- 239000006185 dispersion Substances 0.000 claims description 2
- 230000003993 interaction Effects 0.000 claims description 2
- 238000012847 principal component analysis method Methods 0.000 claims description 2
- 238000011946 reduction process Methods 0.000 claims 1
- 238000005192 partition Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 9
- 239000000243 solution Substances 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 238000010276 construction Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000011217 control strategy Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000013016 damping Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000013210 evaluation model Methods 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000005311 autocorrelation function Methods 0.000 description 1
- 239000003637 basic solution Substances 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000007620 mathematical function Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000005086 pumping Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000010206 sensitivity analysis Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/08—Computing arrangements based on specific mathematical models using chaos models or non-linear system models
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- 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]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Optimization (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Nonlinear Science (AREA)
- Algebra (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Power Engineering (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (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
The invention provides a voltage sag estimation method, medium and system in a large-scale power grid, belonging to the technical field of voltage sag estimation, wherein the method comprises the following steps: collecting power grid data and voltage sag data in a large-scale power grid according to load areas in a partition mode; establishing a simulated dynamics model according to the power grid data, and searching bifurcation points according to the power grid data by using a chaos theory algorithm, wherein the bifurcation points are used for representing the cause points for causing the voltage sag of the power grid; performing evolution fitting on the simulated dynamics model by using all power grid data and voltage sag data; performing thinning treatment on the collected power grid data to obtain thinned power grid data; performing dimension reduction treatment on the evolutionarily fitted simulated dynamics model by using the sparse grid data to obtain a voltage sag assessment model; and estimating the region of the voltage sag which occurs according to the power grid data of the multiple complex regions of the time period to be analyzed by using the voltage sag estimation model, and outputting the estimated region to staff. The technical problem of fast predicting the voltage sag area in real time is solved.
Description
Technical Field
The invention belongs to the technical field of voltage sag estimation, and particularly relates to a voltage sag estimation method, medium and system in a large-scale power grid.
Background
With the development of society, the power system is continuously developed towards more complex and larger scale, and a great challenge is brought to the safe and stable operation of the power system. The voltage sag accident occurs, which seriously affects the economic operation of the power system. Therefore, the voltage sag prediction and evaluation technology is researched, the safe, economical and clean operation of the power system is realized, and the method has great significance for the development of the power industry.
At present, the voltage sag prediction technology mainly comprises the following steps:
1. prediction techniques based on extensive statistical analysis. According to historical data, the technology adopts some simple statistical analysis models to establish the correlation between the voltage sag quantity and influence factors and performs rough voltage sag estimation. This approach predicts less accurately.
2. Prediction techniques based on theoretical analysis modeling. The technology utilizes a mathematical model of the power system to simulate the steady state and dynamic characteristics of the power grid under different running conditions and predict the voltage sag. The method needs detailed system parameters, and has large calculation amount and poor real-time performance.
3. Prediction technology based on artificial intelligence algorithm. The technology applies some intelligent algorithms, such as a neural network, a support vector machine and the like to predict the voltage sag. The method relies on a large amount of historical data to perform model training, and has poor prediction effect on small sample data.
4. Prediction techniques based on probability statistics. This technique treats the voltage dip as a random event and utilizes probabilistic statistical modeling for voltage dip assessment. The influence factors of the method on the voltage sag are not fully considered.
5. Prediction techniques based on multi-time scale cascading. The technology comprehensively considers multiple time scale factors from long term to short term, establishes a layered prediction model, and improves prediction accuracy. But the hierarchical model occasionally generates page prediction errors.
6. Prediction techniques based on big data analysis. The technology utilizes a high-speed storage and calculation platform to collect and process a large amount of data in real time to predict the voltage sag. However, the processing and utilization of the data are insufficient, and the prediction effect can be improved.
In summary, the existing voltage sag prediction technology has the problems of low precision, poor real-time performance, dependence on historical data, large calculated amount and the like. Therefore, a novel method capable of rapidly predicting the voltage sag area in real time is developed, and the method has important significance for safe and stable operation of a power grid.
Disclosure of Invention
In view of the above, the invention provides a voltage sag estimation method, medium and system in a large-scale power grid, which can solve the technical problem of rapidly predicting a voltage sag area in real time.
The invention is realized in the following way:
the first aspect of the present invention provides a voltage sag estimation method in a large-scale power grid, including the steps of:
s10, collecting power grid data and voltage sag data in a large-scale power grid according to a load area in a partitioning mode, wherein the power grid data comprise voltage, current, power and frequency in continuous acquisition time periods, and each acquisition time period is 1-15 seconds; the voltage sag data are time, load area, voltage sag degree and voltage sag duration of occurrence of voltage sag;
s20, establishing a simulated dynamics model according to power grid data, wherein the simulated dynamics model is used for searching bifurcation points according to the power grid data by using a chaos theory algorithm and representing cause points for causing voltage sag of the power grid;
s30, performing evolution fitting on the simulated dynamics model by using all power grid data and voltage sag data;
s40, performing thinning treatment on the collected power grid data to obtain thinned power grid data;
s50, performing dimension reduction treatment on the evolutionarily fitted simulated dynamics model by using the sparse grid data to obtain a voltage sag assessment model;
s60, estimating the region of the voltage sag by utilizing the power grid data of the multiple complex regions of the time period to be analyzed by using the voltage sag estimation model, and outputting the estimated region to a worker.
On the basis of the technical scheme, the voltage sag estimation method in the large-scale power grid can be further improved as follows:
the specific steps of establishing the imitation dynamics model according to the power grid data comprise:
according to the actual operation topological structure of the power grid, a node-branch model is established;
determining a state variable according to the power grid data and the voltage sag data;
establishing a differential equation describing the evolution of each state variable;
setting initial conditions and solving differential equations to obtain a simulation dynamics model.
Further, the criteria for building the node-leg model are: taking the whole power grid as a topological graph, taking each transformer substation as a node, taking lines between the nodes as branches, taking a power station high-voltage bus with the largest system capacity as a public sink, and describing the node-branch model by adopting a node admittance matrix.
Further, the simulated kinetic model comprises:
the long-term dynamic model is used for describing the rotation speed movement of the generator, and the time scale is in the second level, and comprises a rotation speed mechanical equation and a torque load related equation;
the short-term dynamic model has a time scale of millisecond and is used for electromagnetic transient process after network failure, and the interaction and constraint relation of the electric quantity in the system are reflected by adopting differential equations of each order;
the ultra-short term dynamic model has a time scale of microsecond and is used for describing a line and reflecting the dispersion characteristic of the line parameters.
The step of performing evolution fitting on the simulated dynamics model by using all power grid data and voltage sag data specifically comprises the following steps:
defining an objective function of deviation between the output of the simulated dynamics model and measured data, and recording the objective function as a deviation objective function;
finding out an optimal parameter solution of the deviation objective function by using an optimizing algorithm;
the optimal parameter solution iterative optimization is utilized to enable the simulated dynamics model to output approximate actual measurement data;
and (5) obtaining an accurately predicted imitation dynamics model after iteration.
The step of performing thinning treatment on the collected power grid data to obtain the thinned power grid data specifically comprises the following steps:
determining the rarefaction proportion according to the time correlation of the power grid data;
performing equidistant thinning on the power grid data according to the thinning proportion;
and carrying out smoothing treatment on the data obtained by the thinning as the data of the thinning power grid.
The step of performing dimension reduction processing on the evolutionarily fitted simulated dynamics model by utilizing the diluted power grid data to obtain a voltage sag evaluation model specifically comprises the following steps:
extracting main features of the evolutionarily fitted imitation dynamics model by using a principal component analysis method;
constructing a transformation function of mapping the original features to the main components to realize dimension reduction of the simulated dynamics model;
the simulated kinetic model is reconstructed in a low-dimensional space.
The step of predicting the occurred voltage sag data by utilizing the voltage sag evaluation model to estimate the power grid data of multiple complex areas of the time period to be analyzed specifically comprises the following steps:
parallel simulation is carried out in the simulation dynamics model by utilizing the power grid data of multiple areas;
comparing according to the model simulation result to obtain an abnormal region;
and estimating the probability of voltage sag of the abnormal region according to the characteristics of the abnormal region, and outputting the probability to staff.
A second aspect of the present invention provides a computer readable storage medium, where the computer readable storage medium stores program instructions, where the program instructions are executed to perform the method for estimating a voltage sag in a large-scale power grid as described above.
A third aspect of the present invention provides a voltage sag estimation system in a large-scale power grid, wherein the system includes the computer readable storage medium.
Compared with the prior art, the voltage sag estimation method, medium and system in the large-scale power grid provided by the invention have the beneficial effects that: by constructing the simulation dynamics model, the physical characteristics of the power grid can be fully considered, the actual running state of the power grid can be simulated, and the prediction of the dynamic behavior of the power grid can be realized. The real-time prediction of small sample and even single sample faults is realized by combining the model and the data without depending on a large amount of historical data. By utilizing the thinning processing mode, a large amount of historical data is not needed, meanwhile, the complexity of a model is further reduced, the calculated amount of the model is reduced, the calculated speed of the model is improved, and the technical problem of rapidly predicting a voltage sag area in real time is solved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a voltage sag estimation method in a large-scale power grid provided by the invention.
Detailed Description
As shown in fig. 1, an embodiment of a voltage sag estimation method in a large-scale power grid according to a first aspect of the present invention includes the following steps:
s10, collecting power grid data and voltage sag data in a large-scale power grid according to load areas in a partitioning mode, wherein the power grid data comprise voltage, current, power and frequency in continuous acquisition time periods, and each acquisition time period is 1-15 seconds; the voltage sag data are time, load area, voltage sag degree and voltage sag duration of the voltage sag;
s20, establishing a simulated dynamics model according to power grid data, wherein the simulated dynamics model is used for searching bifurcation points according to the power grid data by using a chaos theory algorithm and used for representing cause points causing voltage sag of the power grid;
s30, performing evolution fitting on the simulated dynamics model by using all power grid data and voltage sag data;
s40, performing thinning treatment on the collected power grid data to obtain thinned power grid data;
s50, performing dimension reduction treatment on the evolutionarily fitted simulated dynamics model by using the sparse grid data to obtain a voltage sag assessment model;
s60, estimating the region of the voltage sag by utilizing the power grid data of the multiple complex regions of the time period to be analyzed by using the voltage sag estimation model, and outputting the estimated region to a worker.
The technical principle of the step S10 is that scientific partitioning is carried out aiming at the power grid topology, and the partition simulation model is facilitated to be constructed. The important nodes are selected as characteristic nodes, so that the sampling is more representative. The coverage of the acquisition parameters is wide, and the node state can be fully reflected. And improving the time resolution and acquiring the transient characteristics of the power grid. The dip data is labeled for model training and prediction. The S10 has the technical effect that the data collection is highly targeted according to the regional division of the actual topological condition of the power grid. And selecting characteristic nodes for key monitoring, and ensuring data of the structure sensitive parts of the collected samples. And collecting abundant electrical parameters and comprehensively reflecting the running state of the node. The time resolution is high, and the transient characteristics of the power grid can be known. And the collection of the dip data is matched with the power grid data, so that the samples have pertinence.
The technical principle of the step S20 is that a state space method is utilized to establish a power grid dynamics model. The parameters take measured data, so that the model has real physical meaning. The bifurcation point of the analytical model determines the cause of the fault. And adjusting the model parameters to match the actual fault points. The S20 has the technical effects that a simulation dynamics model is established, and the actual physical process of the power grid is simulated. The model parameters are coupled with the actual high, and the accurate prediction capability is achieved. The cause point of the fault can be found out, and the fault prediction is realized.
The technical principle of the step S30 is to define a fitness function as an optimization target. The optimization algorithm explores the model parameter solutions that minimize the fitness function. And gradually enabling the model output to approach to the actual measurement data through iteration. A high-precision simulation model is obtained. The S30 step has the technical effect that an objective function of deviation between model output and measured data is constructed. And (5) adopting an optimization algorithm to iteratively find an optimal parameter solution. The model can be accurately matched with historical data after optimization. An accurately predicted simulation model is obtained.
The technical principle of the S40 step is that the time correlation is analyzed, and the pumping ratio is determined. And reducing the data volume by adopting an equidistant thinning method. And the thinning data is smoothed, so that the quality is improved. Analysis of feature retention ensures that primary information is contained. The technical effect of step S40 is to reduce the amount of raw data and reduce the amount of storage and computation. The primary data features are retained by analysis. Smoothing reduces the thinning error. And the problem of overfitting caused by big data is avoided.
The technical principle of the step S50 is that the principal component analysis obtains the principal characteristic direction. And constructing linear transformation to realize dimension reduction mapping. The kinetic model is reconstructed in a low dimensional space. And simulating the low-dimensional model to obtain a result. The S50 has the technical effects that main components are analyzed and extracted to obtain main features, and redundancy is filtered. The mapping function implements the transformation of the raw data into a low dimension. The low-dimensional model reduces computational complexity. Preventing the high-dimensional data from overfitting.
The technical principle of the step S60 is that real-time data is input to run a simulation model. And comparing the outputs of the multi-region models to find out fault points. And evaluating the fault situation according to the model result. And (5) deriving a control strategy, and updating the model to form a closed loop. The S60 step has the technical effects that the multi-region model carries out parallel simulation, and the efficiency is improved. And outputting abnormal points by the detection model, and positioning a fault area. The effect of the fault is evaluated and a control reference is provided. An on-line monitoring system is built to realize automatic early warning.
The following describes the specific embodiments of the above steps:
the specific embodiment of step S10 is as follows:
dividing the power grid into a plurality of load areas Z according to the structure and the operation condition of the power grid i (i=1, 2,., n) each load zone contains node information, line parameter information within that zone. The node information comprises node numbers, node types, load amounts and the like; the line parameter information includes line number, line length, line parameter, and the like.
In each load zone Z i And selecting a plurality of characteristic nodes, wherein the characteristic nodes comprise important load nodes of the area, important nodes with poor bearing voltage quality and the like.
And continuously collecting the voltage U, the current I, the active power P and the frequency f for each characteristic node in the collection time period. The acquisition time period is set to be Δt, that is, data is acquired every Δt seconds for each feature node. The collected data for each feature node may be expressed as:
wherein t is n N Δt, n=1, 2. k represents region Z i Is the kth feature node of (c).
Collecting voltage sag data of each load area in the acquisition time period, and recording the occurrence time t s Load region number i, voltage dip size ΔU, voltage dip duration Δt s . The voltage sag data can be expressed as:
VD j :t s ,i,ΔU,Δt s ；
where j represents the jth voltage sag event.
And matching the acquired data of all the characteristic nodes with the voltage sag data to obtain the power grid data of the corresponding characteristic nodes when each voltage sag event occurs. The matching mode is as follows:
for the jth voltage sag event, finding a load area i where voltage sag occurs, and then finding the closest t moment in the acquired data of the characteristic nodes of the area s Acquisition Point t of (2) m And extracting the data of the acquisition point and the front and rear L acquisition points as the data of the sag event. Namely:
matched represents a match.
In conclusion, the step S10 realizes the partition division of the power grid, selects characteristic nodes, collects and matches data, and provides comprehensive power grid data for the subsequent establishment of the model.
In step S20, a specific implementation manner of the simulation dynamics model is established:
firstly, constructing a power grid topology model
Firstly, according to the actual operation topological structure of the power grid, a node-branch model is established. Each bus is taken as a node, and the connecting lines among the nodes are taken as branches. The bus node attributes comprise voltage amplitude, voltage phase angle, active and reactive loads and the like; the properties of the line branches include parameters such as resistance, reactance, susceptance, etc.
(II) determining state variables
According to electrical knowledge, at least the voltage amplitude and the voltage phase angle need to be selected as state variables to describe the dynamic behavior of the power grid. If a finer model needs to be established, active and reactive currents can be injected into the nodes to serve as state variables.
(III) construction of a State equation
Based on node analysis theory, differential equations describing the evolution of each state variable can be established, for example, the voltage phase angle evolution equation of node i can be expressed as:
wherein omega 0 For nominal frequency, θ j K being the voltage phase angle of the other node connected to node i ij And the coupling coefficient between the nodes is obtained by calculation according to the line parameters.
And establishing differential equations describing the evolution of all the state variables, namely a state equation set of the system.
(IV) determining model parameters
The parameters of the model mainly include line parameters and load parameters. The line parameters can be directly obtained from the distribution network parameters, and the load parameters can be estimated through historical load data fitting.
(V) setting initial conditions
Setting the starting simulation time of the model to t 0 At this time, each state variable takes an actual initial value, such as a node voltage phase angle θ i (t 0 ) Taking an actual measurement value.
(VI) solving the equation of state
Solving a state space equation set by adopting a numerical integration algorithm and using a smaller step length to obtain each state variable from t 0 Evolution curves at later moments.
(seventh) model verification
Collecting the time t of failure s Checking whether the simulation result is abnormal, if not, adjusting the model parameters, and repeating the process until the simulation result can reflect the running state of the power grid when the fault occurs.
(eight) optionally, analyzing and applying
Based on the verified model, different control measures can be replaced, the fault evolution condition is observed, and the optimal state control strategy is found out. The initial conditions may also be modified for contextual analysis.
The specific steps of the above (one) to (eight) are described in detail below:
the construction of the power grid topology model is described in detail as follows:
1. collecting grid configuration data
And collecting configuration data of the power grid, including information such as transformer substations, lines, loads and the like. The main data include:
(1) Substation data including substation numbers, names, voltage levels and the like;
(2) Line data, including line start point, line end point, resistor r, reactance x, susceptance b and other parameters;
(3) Load data, load number, bus, load power P (megawatt), load power factorEtc.
2. Establishing node-branch model
The collected data are consolidated into a node-branch model:
(1) Each transformer station is taken as a node;
(2) The circuit between the nodes is taken as a branch circuit;
(3) The branch parameters comprise a starting point, an ending point, a resistor r, a reactance x and a susceptance b;
(4) The node attributes include voltage magnitude U, phase angle θ, and connected leg information.
The entire grid can be modeled as one graph g= (V, E):
v is a node set and represents each transformer substation;
e is a branch set, representing the line between nodes.
3. Handling special nodes
For the generator node, the phase electromotive force E and the internal resistance r need to be increased a Isoparametric parameters; for a load node, the load power P and the power factor of the load node need to be determined
4. Determining a public sink
The public sink of the system is selected, and the high-voltage bus of the power station with the maximum system capacity is generally selected. The sink voltage magnitude U and phase angle θ are taken as references, and the other node voltages are relative to the sink.
5. Establishing node admittance matrix
For the above node-branch model, the node admittance matrix Y may be established by a node analysis method.
6. Summarizing into a power grid topology model
So far, we have built a detailed topology model of the grid, described as follows:
node set V representing each substation
Branch set E representing inter-node lines
Node admittance matrix Y describing network connection relationship
Attributes such as voltage U, phase angle theta and the like of each node
The topology model lays a foundation for subsequent power grid analysis modeling.
(II) determining state variables is elaborated as follows:
the kinematic characteristics of the power grid to be described are comprehensively considered, and at least the voltage amplitude U and the voltage phase angle theta are required to be selected as state variables so as to describe the motion state of the power grid.
According to the model precision requirement, the active injection current I of the node can be obtained P And reactive injection current I Q Also referred to as state variables, to build a finer model.
If the motor motion equation is taken into account, the generator rotational speed deviation Δω and the rotor angular deviation Δδ can also be selected as state variables.
For a power network comprising converter devices, the DC side voltage U can be increased dc As state variables.
(III) constructing a state equation, and elaborating as follows:
for each state variable x i From electrical knowledge, a differential equation describing its dynamics is established:
for example, the voltage phase angle dynamics equation can be expressed as:
wherein K is ij Is the coupling coefficient between the nodes.
The state equation fully describes the dynamic evolution relation of each state variable, and lays a foundation for building a power grid simulation model.
(IV) determining model parameters, wherein the model parameters are described in detail as follows:
the parameters of the power grid model mainly comprise node parameters, branch parameters and equipment parameters.
The node parameters include load power P Lj Power factorEtc.:
the branch parameter includes a resistor r l Reactance x l Susceptance b l :
Line l :[r l ,x j ,b j ,...]；
The plant parameters include generator parameters and parameters of other plants.
The generator parameters include moment of inertia M in the equation of motion i Damping coefficient D i Etc.; other devices include parameters of transformers.
The parameters can be obtained from a power grid configuration database, or can be estimated by a parameter estimation algorithm by using historical operation data.
Parameters can also be learned from big data through a data driving method and a deep learning algorithm.
The parameter determination process comprises the following steps:
(1) Collecting required parameter information
(2) Estimating missing parameters
(3) Applying parameter estimation and data learning algorithms
(4) Obtaining full parameter set of power grid model
(V) setting initial conditions, which are explained in detail as follows:
the initial conditions of the model correspond to the initial time t 0 State variable values of (2).
Initial voltage amplitude U j0 ＝U j (t 0 ) Initial phase angle θ j0 ＝θ j (t 0 ) Can be obtained from measurements.
Initial rotational speed deviation Δω i0 ＝Δω i (t 0 ) Taken as 0.
Initial phase angle deviation delta i0 ＝Δδ i (t 0 ) Taken as 0.
And the initial values of other state variables are measured if the state variables can be measured, and reasonable estimated values are obtained if the state variables cannot be measured.
The overall initial condition may be expressed as an initial vector of state variables:
X 0 ＝[U 10 ,θ 10 ,...,Δω 10 ,Δδ 10 ,...]；
the process of setting the initial conditions is:
(1) Collecting initial values of measurable state variables
(2) Giving reasonable estimates of non-measurable variables
(3) Summarizing into initial condition vectors
And (six) solving a state equation, wherein the state equation is described in detail as follows:
the system of state equations forms a rigid nonlinear differential equation system that needs to be solved by a numerical method.
The most basic solution is the euler forward difference method.
And repeating the steps to obtain a state evolution curve of the system in the whole time domain.
The higher-precision solving method comprises a Runge-Kutta method and the like, and can improve the accuracy of numerical calculation.
The solving process needs to write a program, and a numerical algorithm is implemented on a computer to obtain a time response result of the state variable.
(seventh) optionally, model verification, elaborating as follows:
collecting power grid data when actual faults occur, and obtaining fault time t s Is set, the value of each state variable of (a).
On the simulation model, given the same initial conditions, run to t s Time of day.
And comparing the deviation of the simulation result and the actual measured value.
If the deviation is too large, the parameters of the model need to be adjusted, or the form of the state equation is modified, and then the above-mentioned process is repeated.
Model verification is considered to be passed when the simulation results are able to match the measured values accurately.
(eight) optionally, model application, elaborated on as follows:
on the validated model, various simulation analyses can be performed.
And (3) fault evolution simulation, namely predicting a fault evolution process and a final state in the initial stage of the fault.
And (3) fault control simulation, namely, assuming different control measures to be adopted, analyzing the dynamic response of the power grid.
And (3) parameter sensitivity analysis, namely changing parameters and observing the change condition of the power grid movement.
Through simulation test, a basis is provided for operation control of the power grid.
The specific embodiment of step S30 is as follows:
construction of model fitness function
A mathematical function is defined to measure the degree of fit between the model output and the actual data. For example, root mean square error may be employed:
wherein y is i Is the actual data that is to be used,is the model output and n is the number of samples. Smaller F values indicate better fitting.
(II) determining model adjustable parameters
Many undetermined parameters are included in the simulation model and need to be optimized by fitting. These parameters are determined as decision variables in the optimization.
For example, the generator damping coefficient D may be used as an adjustable parameter.
A search algorithm is used to try the parameter combinations to minimize the fitness value F of the model.
For example, genetic algorithms may be used:
(1) Initializing a set of parameter combinations;
(2) Calculating the F value of the model under each group of parameters;
(3) The combination with lower F value (good fitting effect) is reserved, and the combination with higher F value is eliminated;
(4) Generating new parameter combinations through crossover, mutation and other operations;
(5) Returning to the step (2) until an acceptable minimum F value is obtained.
And repeating the trial optimizing process until the model output can highly fit the actual data.
And obtaining a group of optimization parameters which enable the model to achieve a good fitting effect.
And redefining a simulation model by using the optimized parameters to complete the evolution fitting optimization of the model.
The specific embodiment of step S40 is as follows:
analysis of time series characteristics of raw grid data
The collected power grid simulation data comprise a large number of high-frequency samples, and the data volume is reduced by performing thinning. Before the thinning, the time correlation of the original data is firstly analyzed, and the acceptable thinning proportion is determined.
For example, the autocorrelation function method is adopted to analyze the correlation degree of the sampled data in time, and if the correlation coefficient is fast decayed after a certain time lag tau, the validity of the data can be ensured by sampling every tau sampling points.
(II) determining the thinning proportion
And determining the thinning proportion p of the original data according to the time sequence analysis. For example, it is determined to extract one sampling point every p points. The thinning proportion should not be too large so as not to lose excessive information.
Downsampling data by using equidistant thinning method
Traversing the original data, extracting a point every p points, and reserving the sampling points to form a thinned data set.
For example, the original data is x 1 ,x 2 ,...,x n ；
The data set after thinning is x 1 ,x 1+p ,x 1+2p ,...；
(III) smoothing the thinned data set
Because the equal interval thinning may cause data fluctuation, the thinning data set can be smoothed by adopting a moving average technology and the like so as to reduce errors caused by thinning.
And analyzing the data after the thinning to ensure that the main characteristics of the original data are reserved.
For example, comparing the characteristic values, spectrum contents, etc. of the original data and the thinned data, if the main characteristics remain consistent, it is reasonable to indicate the thinning method.
And using the thinned data set as the input of the simulation model to replace the original large data set.
The specific embodiment of step S50 is as follows:
analyzing the model data characteristics to determine the principal components
Assuming that the data sample matrix of the simulation model is X m×n Where m is the number of samples and n is the number of features.
(1) Calculating a covariance matrix C of the sample data matrix X:
(2) Performing eigenvalue decomposition on the covariance matrix C:
C＝VΛV T ；
wherein Λ is a diagonal matrix, and diagonal elements are eigenvalues λ i The method comprises the steps of carrying out a first treatment on the surface of the V is the corresponding eigenvector matrix.
(3) Sorting according to the order of the eigenvalues, selecting the eigenvectors corresponding to the first k eigenvalues to form a principal component transformation matrix P:
P＝[v 1 ,v 2 ,...,v k ]；
wherein the ith feature vector v i Corresponding characteristic value lambda i 。
(4) The k principal components determined reflect the principal features in the original data.
Establishing a transformation of original features into principal components
A principal component transformation matrix P has been obtained, the column vector of which is the direction of each principal component.
For any original data sample x, the principal component representation z is:
z＝P T x；
the matrix P gives a transformation from the original feature space to the low-dimensional principal component space.
Projection of data into a low dimensional space
For each data sample x generated by the simulation model, the principal component representation is calculated by applying the above-mentioned transformation:
z i ＝P T x i ,i＝1,2,...,m；
i.e. a projection of the raw data into a low-dimensional space is achieved.
(II) reconstructing a low-dimensional simulation model
In the low-dimensional principal component space, a new state equation is constructed:
wherein, the state variable z is vector after dimension reduction, and the input u' also reduces dimension.
Function f ′ And a modeling method similar to the original model is adopted, so that the physical characteristics under the representation of the principal component are satisfied.
(III) solving the low-dimensional model
And applying the same numerical solution algorithm as the original model to the new low-dimensional model to obtain a dynamic evolution curve of the dimension reduction state z.
Wherein principal component analysis produces a plurality of principal components, each principal component corresponding to one orthogonal direction of the sample dataset. The principal components are arranged in order of magnitude representing the variance of the sample, the first principal component representing the direction of maximum change of the sample, the second principal component representing the direction of second maximum change of the sample, and so on.
In selecting the principal component, there are two common methods:
1. by eigenvalue method
The characteristic value is selected according to the magnitude of the characteristic value corresponding to the main component, and the larger the characteristic value is, the larger the sample change reflected by the main component is. The first few principal components are typically selected for which the eigenvalue cumulative variance contribution reaches a certain threshold (e.g., 80%).
2. By cutting-off
The number of selected principal components is specified in advance, for example, the first m principal components are selected. m is selected according to the characteristics of the sample data, and the number of main components capable of representing the main change direction of the sample is generally selected.
In addition to the above method, selection can also be made by observing the specific directional meaning of the principal component. For example, a principal component load matrix may be drawn, and the contribution of each principal component to the original variable is analyzed, and principal components that are more explanatory of the sample are selected.
In general, the selection of principal components is required in conjunction with specific data sets and research objectives. The purpose is to select the minimum number of principal components that can represent the principal characteristics of the sample, both to reduce redundant information and to preserve the integrity of the original information.
The specific embodiment of step S60 is as follows:
collecting real-time power grid data of each area of a time period to be analyzed
Including information on voltage, current, power, etc., and performs necessary preprocessing such as denoising, normalization, etc.
(II) inputting real-time data and running a multi-region simulation model
And inputting the preprocessed real-time power grid data into simulation models of different areas, starting and running the models to obtain simulation results of all state variables.
(III) detecting whether abnormal characteristics appear in the model output
The simulation results are analyzed to check whether characteristics indicating voltage sag, such as voltage sag, instability, etc., are present. An appropriate failure determination threshold may be set.
(IV) comprehensively judging the area possibly suffering from faults
And comparing the output of the different region models, and comprehensively judging the region or node most likely to generate voltage sag fault.
(V) optionally, evaluating the possible impact of the fault
If it is determined that there is evidence of failure, the extent of possible failure, the extent of influence, etc. are evaluated based on the simulation model.
In the above steps, the calculation methods such as denoising, normalization, similarity and the like may be conventional methods, and will not be described in detail.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the invention is subject to the protection scope of the claims.
Claims (10)
1. The voltage sag estimation method in the large-scale power grid is characterized by comprising the following steps of:
s10, collecting power grid data and voltage sag data in a large-scale power grid according to a load area in a partitioning mode, wherein the power grid data comprise voltage, current, power and frequency in continuous acquisition time periods, and each acquisition time period is 1-15 seconds; the voltage sag data are time, load area, voltage sag degree and voltage sag duration of occurrence of voltage sag;
s20, establishing a simulated dynamics model according to power grid data, wherein the simulated dynamics model is used for searching bifurcation points according to the power grid data by using a chaos theory algorithm and representing cause points for causing voltage sag of the power grid;
s30, performing evolution fitting on the simulated dynamics model by using all power grid data and voltage sag data;
s40, performing thinning treatment on the collected power grid data to obtain thinned power grid data;
s50, performing dimension reduction treatment on the evolutionarily fitted simulated dynamics model by using the sparse grid data to obtain a voltage sag assessment model;
s60, estimating the region of the voltage sag by utilizing the power grid data of the multiple complex regions of the time period to be analyzed by using the voltage sag estimation model, and outputting the estimated region to a worker.
2. The method for estimating voltage dip in a large-scale power grid according to claim 1, wherein the specific step of establishing the simulated dynamics model according to the power grid data comprises:
according to the actual operation topological structure of the power grid, a node-branch model is established;
determining a state variable according to the power grid data and the voltage sag data;
establishing a differential equation describing the evolution of each state variable;
setting initial conditions and solving differential equations to obtain a simulation dynamics model.
3. The method for estimating voltage dip in a large-scale power grid according to claim 2, wherein the criteria for establishing the node-branch model are: taking the whole power grid as a topological graph, taking each transformer substation as a node, taking lines between the nodes as branches, taking a power station high-voltage bus with the largest system capacity as a public sink, and describing the node-branch model by adopting a node admittance matrix.
4. The method for estimating voltage dip in a large-scale power grid according to claim 2, wherein said simulated dynamics model comprises:
the long-term dynamic model is used for describing the rotation speed movement of the generator, and the time scale is in the second level, and comprises a rotation speed mechanical equation and a torque load related equation;
the short-term dynamic model has a time scale of millisecond and is used for electromagnetic transient process after network failure, and the interaction and constraint relation of the electric quantity in the system are reflected by adopting differential equations of each order;
the ultra-short term dynamic model has a time scale of microsecond and is used for describing a line and reflecting the dispersion characteristic of the line parameters.
5. The method for estimating voltage dip in a large-scale power grid according to claim 1, wherein said step of performing evolution fitting on the simulated dynamics model using all power grid data and voltage dip data comprises:
defining an objective function of deviation between the output of the simulated dynamics model and measured data, and recording the objective function as a deviation objective function;
finding out an optimal parameter solution of the deviation objective function by using an optimizing algorithm;
the optimal parameter solution iterative optimization is utilized to enable the simulated dynamics model to output approximate actual measurement data;
and (5) obtaining an accurately predicted imitation dynamics model after iteration.
6. The method for estimating voltage dip in a large-scale power grid according to claim 1, wherein the step of performing thinning processing on the collected power grid data to obtain thinned power grid data specifically comprises:
determining the rarefaction proportion according to the time correlation of the power grid data;
performing equidistant thinning on the power grid data according to the thinning proportion;
and carrying out smoothing treatment on the data obtained by the thinning as the data of the thinning power grid.
7. The method for estimating voltage dip in a large-scale power grid according to claim 1, wherein the step of performing a dimension reduction process on the evolutionarily fitted simulated dynamics model by using the sparse power grid data to obtain a voltage dip estimation model specifically comprises the following steps:
extracting main features of the evolutionarily fitted imitation dynamics model by using a principal component analysis method;
constructing a transformation function of mapping the original features to the main components to realize dimension reduction of the simulated dynamics model;
the simulated kinetic model is reconstructed in a low-dimensional space.
8. The method for estimating voltage dip in a large-scale power grid according to claim 1, wherein the step of estimating the voltage dip data by using the voltage dip estimation model for the power grid data of the multiple complex areas of the time period to be analyzed specifically comprises:
parallel simulation is carried out in the simulation dynamics model by utilizing the power grid data of multiple areas;
comparing according to the model simulation result to obtain an abnormal region;
and estimating the probability of voltage sag of the abnormal region according to the characteristics of the abnormal region, and outputting the probability to staff.
9. A computer readable storage medium, wherein program instructions are stored in the computer readable storage medium, which program instructions, when executed, are adapted to perform the method for estimating voltage sag in a large-scale electrical network according to any one of claims 1-8.
10. A voltage sag estimation system in a large-scale power grid, comprising the computer-readable storage medium of claim 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311399848.1A CN117439068B (en) | 2023-10-26 | Voltage sag estimation method, medium and system in large-scale power grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311399848.1A CN117439068B (en) | 2023-10-26 | Voltage sag estimation method, medium and system in large-scale power grid |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117439068A true CN117439068A (en) | 2024-01-23 |
| CN117439068B CN117439068B (en) | 2024-05-14 |
Family
ID=
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060111860A1 (en) * | 2002-11-06 | 2006-05-25 | Aplicaciones En Informatica Avanzada, S.A. | System and method for monitoring and managing electrical power transmission and distribution networks |
| WO2016026355A1 (en) * | 2014-08-18 | 2016-02-25 | 国家电网公司 | Voltage sag simulation and evaluation method of active power distribution grid |
| CN106372812A (en) * | 2016-09-23 | 2017-02-01 | 广东工业大学 | Voltage sag risk assessment method and system |
| CN107769197A (en) * | 2017-11-14 | 2018-03-06 | 国网江苏省电力公司电力科学研究院 | A kind of voltage sag source alignment system based on grid equipment topology |
| CN114021424A (en) * | 2021-09-29 | 2022-02-08 | 国网江苏省电力有限公司南京供电分公司 | PCA-CNN-LVQ-based voltage sag source identification method |
| CN115128395A (en) * | 2022-06-14 | 2022-09-30 | 武汉大学 | Voltage sag source positioning method and device |
| CN115877068A (en) * | 2022-06-29 | 2023-03-31 | 国网山西省电力公司电力科学研究院 | Voltage sag propagation track identification method of regional power grid based on deep learning |
| CN115907526A (en) * | 2022-11-11 | 2023-04-04 | 国网电力科学研究院有限公司 | Power grid voltage sag evaluation method and system, storage medium and computing device |
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060111860A1 (en) * | 2002-11-06 | 2006-05-25 | Aplicaciones En Informatica Avanzada, S.A. | System and method for monitoring and managing electrical power transmission and distribution networks |
| WO2016026355A1 (en) * | 2014-08-18 | 2016-02-25 | 国家电网公司 | Voltage sag simulation and evaluation method of active power distribution grid |
| CN106372812A (en) * | 2016-09-23 | 2017-02-01 | 广东工业大学 | Voltage sag risk assessment method and system |
| CN107769197A (en) * | 2017-11-14 | 2018-03-06 | 国网江苏省电力公司电力科学研究院 | A kind of voltage sag source alignment system based on grid equipment topology |
| CN114021424A (en) * | 2021-09-29 | 2022-02-08 | 国网江苏省电力有限公司南京供电分公司 | PCA-CNN-LVQ-based voltage sag source identification method |
| CN115128395A (en) * | 2022-06-14 | 2022-09-30 | 武汉大学 | Voltage sag source positioning method and device |
| CN115877068A (en) * | 2022-06-29 | 2023-03-31 | 国网山西省电力公司电力科学研究院 | Voltage sag propagation track identification method of regional power grid based on deep learning |
| CN115907526A (en) * | 2022-11-11 | 2023-04-04 | 国网电力科学研究院有限公司 | Power grid voltage sag evaluation method and system, storage medium and computing device |
Non-Patent Citations (2)
| Title |
|---|
| HANG QI ETAL.: "Stacking-based estimation of maximal transient voltage drop with unified fault location representation", 2020 IEEE SUSTAINABLE POWER AND ENERGY CONFERENCE (ISPEC), 18 February 2021 (2021-02-18), pages 2391 - 2397 * |
| 李妍, 余欣梅, 熊信艮, 段献忠: "电力系统电压暂降分析计算方法综述", 电网技术, vol. 28, no. 14, 31 July 2004 (2004-07-31), pages 74 - 78 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102074955B (en) | Method based on knowledge discovery technology for stability assessment and control of electric system | |
| Noureen et al. | Analysis and application of seasonal ARIMA model in Energy Demand Forecasting: A case study of small scale agricultural load | |
| CN108155648A (en) | Method for estimating state based on the infinite Extended Kalman filter of adaptive H | |
| CN111523778A (en) | Power grid operation safety assessment method based on particle swarm algorithm and gradient lifting tree | |
| CN111401749A (en) | Dynamic safety assessment method based on random forest and extreme learning regression | |
| CN103778305B (en) | k-VNN- and LS-SVM-based modelling method for icing of electric transmission line | |
| CN111628494B (en) | Low-voltage distribution network topology identification method and system based on logistic regression method | |
| CN104578053A (en) | Power system transient stability prediction method based on disturbance voltage trajectory cluster features | |
| CN112069727A (en) | Intelligent transient stability evaluation system and method with high reliability for power system | |
| CN111525587A (en) | Reactive load situation-based power grid reactive voltage control method and system | |
| CN114595858A (en) | Short-term wind speed prediction method and system based on rolling time series and support vector machine | |
| CN117439068B (en) | Voltage sag estimation method, medium and system in large-scale power grid | |
| CN105260521B (en) | A kind of load Fusion Modeling Method based on analytic hierarchy process (AHP) | |
| CN112101673A (en) | Power grid development trend prediction method and system based on hidden Markov model | |
| CN117439068A (en) | Voltage sag estimation method, medium and system in large-scale power grid | |
| CN115267409A (en) | Power equipment fault detection method, device and equipment | |
| Hu et al. | Power grid's Intelligent Stability Analysis based on big data technology | |
| CN109842113B (en) | Power system simplified equivalence method based on generator group dynamic feature analysis | |
| Shi et al. | Study on stability feature extraction of power system using deep learning | |
| Dabou et al. | Supervised learning of overcomplete dictionaries for rapid response-based dynamic stability prediction | |
| Wang et al. | Ensemble Learning Model of Power System Transient Stability Assessment Based on Bayesian Model Averaging Method | |
| Basia et al. | Comparison of data driven algorithms for SoH estimation of Lithium-ion batteries | |
| CN115965288B (en) | IWO (IWO-based optimization BiLSTM) based active power distribution network frequency security assessment method | |
| Atamuradov et al. | Evaluation of Features with Changing Effectiveness for Prognostics | |
| CN115796336A (en) | Transformer area line loss prediction method based on feature selection and GRU network |
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 |