CN109508830B - Method for predicting space-time dynamic load of electric automobile - Google Patents
Method for predicting space-time dynamic load of electric automobile Download PDFInfo
- Publication number
- CN109508830B CN109508830B CN201811360331.0A CN201811360331A CN109508830B CN 109508830 B CN109508830 B CN 109508830B CN 201811360331 A CN201811360331 A CN 201811360331A CN 109508830 B CN109508830 B CN 109508830B
- Authority
- CN
- China
- Prior art keywords
- dynamic load
- space
- load
- neural network
- matrix
- 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
- 238000000034 method Methods 0.000 title claims abstract description 42
- 239000011159 matrix material Substances 0.000 claims abstract description 57
- 230000001364 causal effect Effects 0.000 claims abstract description 40
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 35
- 238000010606 normalization Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000012360 testing method Methods 0.000 claims abstract description 21
- 238000003062 neural network model Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000012545 processing Methods 0.000 claims abstract description 6
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 239000011800 void material Substances 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000004134 energy conservation Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
- Y04S30/00—Systems supporting specific end-user applications in the sector of transportation
- Y04S30/10—Systems supporting the interoperability of electric or hybrid vehicles
- Y04S30/12—Remote or cooperative charging
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Primary Health Care (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a method for predicting space-time dynamic load of an electric automobile, which comprises the following steps: preprocessing charging pile load data; constructing a space-time dynamic load matrix according to the charging load data and the charging pile position information; carrying out normalization processing on the space-time dynamic load matrix to obtain a space-time dynamic load normalization matrix; dividing the space-time dynamic load normalization matrix into a training set and a test set; training according to the training set to obtain a two-dimensional cavity causal convolutional neural network model; and testing the model, if the parameters in the two-dimensional cavity causal convolutional neural network model enable the target function to be the minimum on the test set, predicting according to the model, and performing inverse normalization, otherwise, adjusting the hyper-parameters of the two-dimensional cavity causal convolutional neural network model, and re-acquiring the two-dimensional cavity causal convolutional neural network model. The two-dimensional cavity causal convolution neural network model can fully consider time and space, and accurate prediction of the space-time dynamic load of the electric automobile is achieved.
Description
Technical Field
The application relates to the technical field of power system operation and load prediction, in particular to a method for predicting a space-time dynamic load of an electric vehicle.
Background
The electric automobile is considered to be one of the beneficial ways for solving the problems of energy shortage and environment because of the characteristics of energy conservation, emission reduction, environmental protection and the like, so that the electric automobile is greatly supported and popularized by governments and enterprises of various countries. Due to uncertainty and mutual difference of electric vehicle user requirements and behaviors, future large-scale electric vehicle charging loads have uncertain characteristics such as randomness, intermittence and volatility in time and space, and bring difficulty to safe operation and optimized scheduling of a power grid, so that the electric vehicle charging loads need to be effectively predicted.
At present, the methods for predicting the load of the electric vehicle mainly comprise two methods. Firstly, a method for predicting the load of the electric automobile by adopting a mathematical model starts from the characteristics of daily driving mileage and daily parking demand space-time distribution of the electric automobile and analyzes the charging demand. Specifically, a Monte Carlo simulation method is adopted to simulate parking, driving and charging behaviors of the electric automobile in different time and space, and the time-space distribution characteristic of the charging load of the electric automobile is predicted. Secondly, a method for predicting by adopting a statistical learning model based on historical data learns the potential law of the historical data by the model, thereby achieving the effect of predicting the time-space distribution characteristic of the charging load of the electric automobile.
In the first method, when the time-space characteristics of the charging load are comprehensively considered, too many factors need to be considered, the mathematical model is too complex, and the prediction precision is difficult to ensure. Therefore, it is necessary to design a method capable of accurately predicting the space-time dynamic load of the electric vehicle.
Disclosure of Invention
The application provides a method for predicting a space-time dynamic load of an electric automobile, which aims to solve the technical problem that the space-time dynamic load of the electric automobile cannot be accurately and effectively predicted in the prior art.
In order to solve the technical problem, the embodiment of the application discloses the following technical scheme:
the embodiment of the application discloses a method for predicting space-time dynamic load of an electric automobile, which comprises the following steps:
step S101: preprocessing charging pile load data;
step S102: constructing a space-time dynamic load matrix according to the charging load data and the charging pile position information;
step S103: carrying out normalization processing on the space-time dynamic load matrix to obtain a space-time dynamic load normalization matrix;
step S104: dividing the space-time dynamic load normalization matrix into a training set and a test set;
step S105: training according to the training set to obtain a two-dimensional cavity causal convolution neural network model;
step S106: testing the result of the two-dimensional cavity causal convolutional neural network model according to the test set, if the parameters in the two-dimensional cavity causal convolutional neural network model enable the target function to be the minimum on the test set, performing step S107, otherwise, adjusting the hyper-parameters of the two-dimensional cavity causal convolutional neural network model, and returning to step S105;
step S107: and predicting according to the two-dimensional cavity causal convolution neural network model, and performing inverse normalization.
Preferably, in the method for predicting the spatiotemporal dynamic load of the electric vehicle, the preprocessing of the charging pile load data includes:
checking missing values and abnormal values in the charging pile load data;
and removing the missing value and the abnormal value, and filling effective values according to a Lagrangian difference value method.
Preferably, in the method for predicting the spatiotemporal dynamic load of the electric vehicle, the constructing a spatiotemporal dynamic load matrix according to the charging load data and the charging pile position information includes:
step S301: establishing coordinate axes, determining the coordinates of the charging piles, and calculating the load coverage range of each charging pile;
step S302: filling the load quantities of the charging piles at the first moment in the load coverage range of all the charging piles and accumulating to obtain a load matrix at the first moment
In the formula (I), the compound is shown in the specification,
is the load quantity of the point with coordinates (x, y);
step S303: repeating the step S302 according to the time sequence until all the moments in the time length T contained in the electric automobile load data construct a two-dimensional load matrix, and arranging a space-time sequence D ═ D { (D) 1 ,D 2 ,...,D T },D∈R T×X×Y 。
Preferably, in the method for predicting a space-time dynamic load of an electric vehicle, the normalizing the space-time dynamic load matrix to obtain a space-time dynamic load normalization matrix includes:
in the space-time dynamic load normalization matrix,
in the formula, X max Is the maximum of all matrix elements, X i Is the i matrix element value.
Preferably, in the method for predicting the spatio-temporal dynamic load of the electric vehicle, the dividing the spatio-temporal dynamic load normalization matrix into a training set and a test set includes: 80% of the space-time dynamic load normalization matrix data are training sets, and 20% of the space-time dynamic load normalization matrix data are training sets.
Preferably, in the method for predicting a space-time dynamic load of an electric vehicle, the training according to the training set to obtain a two-dimensional cavity causal convolutional neural network model includes:
convolving the jth data of the ith layer with x, y and z positions, and rollingThe product result is:
and the convolution kernel is (2 w h), the size r of the receptive field is 2 L- 1 R i Wherein d is 2 l-1 ，R i Is the size of the first dimension of the three-dimensional convolution kernel, and R i ＝2；
And stacking the convolution results in sequence to form the two-dimensional cavity causal convolution neural network model M (·).
Preferably, in the method for predicting the spatiotemporal dynamic load of the electric vehicle, the predicting according to the two-dimensional cavity causal convolutional neural network model includes:
prediction value
Wherein N is r.
Preferably, in the method for predicting the spatio-temporal dynamic load of the electric vehicle, an objective function of the two-dimensional cavity causal convolutional neural network model M (-) is as follows:
wherein W, b is a parameter for predicting the network structure, γ is the weight of the regularization term, D ture Are true values.
Preferably, in the method for predicting the spatiotemporal dynamic load of the electric vehicle, adjusting the hyper-parameter of the two-dimensional cavity causal convolutional neural network model includes:
and obtaining a two-dimensional cavity causal convolution neural network model parameter which enables the objective function to be minimum through an adam random gradient descent method.
Preferably, in the method for predicting the spatiotemporal dynamic load of the electric vehicle, the performing of the inverse normalization includes: x i ＝X′ i ×X max ，
In the formula, X max Is the maximum of all matrix elements, X i Is the i matrix element value.
Compared with the prior art, the beneficial effect of this application is:
the application provides a method for predicting the spatio-temporal dynamic load of an electric automobile, comprising the steps of preprocessing charging pile load data, constructing a spatio-temporal dynamic load matrix according to the charging load data and charging pile position information, normalizing the spatio-temporal dynamic load matrix, dividing the normalized spatio-temporal dynamic load matrix into a training set and a testing set, training according to the training set to obtain a two-dimensional cavity cause and effect convolutional neural network model, testing the result of the two-dimensional cavity cause and effect convolutional neural network model according to the testing set, if the parameters in the two-dimensional cavity cause and effect convolutional neural network model enable an objective function to be minimum on the testing set, predicting according to the two-dimensional cavity cause and effect convolutional neural network model and carrying out inverse normalization, otherwise, adjusting the hyper-parameters of the two-dimensional cavity cause and effect convolutional neural network model, and recalculating the two-dimensional cavity causal convolutional neural network model. The two-dimensional cavity causal convolutional neural network model can learn information of space dimensionality and can receive long-term historical input, so that the two-dimensional cavity causal convolutional neural network model can learn time dimensionality information, and accurate prediction of electric automobile space-time dynamic load is finally achieved.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application.
Drawings
In order to more clearly explain the technical solution of the present application, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for predicting a space-time dynamic load of an electric vehicle according to an embodiment of the present invention;
fig. 2 is a distribution diagram of charging piles according to an embodiment of the present invention;
FIG. 3 is a diagram of a two-dimensional hole convolutional neural network provided by an embodiment of the present invention;
fig. 4 is a diagram of the predicted result provided by the embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, a flow chart of a method for predicting a space-time dynamic load of an electric vehicle according to an embodiment of the present invention is shown. As can be obtained by referring to fig. 1, the method for predicting the spatiotemporal dynamic load of the electric vehicle in the present application includes:
step S101: preprocessing charging pile load data;
and checking a missing value and an abnormal value in the charging pile load data, removing the missing value and the abnormal value, and filling an effective value according to a Lagrangian difference value method. Fill electric pile load data and all probably produce the missing value in collection, transmission, storage process, directly look over can obtain, the abnormal value need combine the judgement through boxplot and actual conditions to obtain, like negative value, too big value and undersize value etc.. The quality of the charging pile load data has great influence on the model prediction accuracy, so that reasonable values need to be refilled into the abnormal values and the missing values in advance, and the abnormal values and the missing values are filled by a Lagrangian difference method.
Step S102: constructing a space-time dynamic load matrix according to the charging load data and the charging pile position information;
the electric vehicle load has randomness in time and space, and in order to better predict the space-time dynamics, the load on the charging pile needs to be depicted in space-time dimension. Referring to fig. 2, in the charging pile distribution diagram provided in the embodiment of the present invention, as shown in fig. 2, a matrix with a length and a width of X, Y is established according to longitude and latitude distribution of 10 charging piles, wherein the size of X, Y is determined according to actual position distribution of the charging piles on the map, and the values selected by X, Y in fig. 2 are 40, and the step of constructing the matrix includes:
step S301: establishing coordinate axes, determining the coordinates of the charging piles, and calculating the load coverage range of each charging pile, wherein the coverage range of each charging pile is an L multiplied by L square with the coordinates of the charging pile as the center;
step S302: filling the load quantities of the charging piles at the first moment in the load coverage range of all the charging piles and accumulating to obtain a load matrix at the first moment
In the formula (I), the compound is shown in the specification,
is the load quantity of the point with coordinates (x, y);
step S303: repeating the step S302 according to the time sequence until all the moments in the time length T contained in the electric automobile load data construct a two-dimensional load matrix, and arranging a space-time sequence D ═ D { (D) 1 ,D 2 ,...,D T },D∈R T×X×Y 。
Step S103: carrying out normalization processing on the space-time dynamic load matrix to obtain a space-time dynamic load normalization matrix;
for the convenience of data processing and the acceleration of the network learning speed, the time-space dynamic load matrix is normalized in the application, in the time-space dynamic load normalization matrix,
in the formula, X max Is the maximum value, X, of all matrix elements i Is the i matrix element value.
Step S104: dividing the space-time dynamic load normalization matrix into a training set and a test set;
the first 80% of the space-time dynamic load normalization matrix data is used as a training set, and the last 20% of the space-time dynamic load normalization matrix data is used as the training set.
Step S105: training according to the training set to obtain a two-dimensional cavity causal convolution neural network model;
the establishment of the charging pile space-time dynamic load matrix for predicting K time points in the future requires the establishment of a cause-and-effect system p for predicting K time points in the future according to observed values of S time points in the past,
the cause-and-effect system p is a two-dimensional lost motion cause-and-effect convolution neural network provided in the present application, and the idea is to combine a three-dimensional convolution structure applied to a space dimension and a one-dimensional void cause-and-effect convolution structure to form the two-dimensional void cause-and-effect convolution neural network, that is, to replace the one-dimensional convolution of the one-dimensional void convolution with a three-dimensional convolution, specifically, the convolution process is as follows:
convolving jth data of the ith layer with x, y and z positions, wherein the convolution result is as follows:
and the convolution kernel is (2 w h), the size r of the receptive field is 2 L- 1 R i Wherein d is 2 l-1 ，R i Is the size of the first dimension of the three-dimensional convolution kernel, and R i And 2, stacking the convolution results in sequence to form the two-dimensional cavity causal convolution neural network model M (·).
Size r of receptive field is 2 L-1 R i In the present application, R is set to indicate that the historical load data of the number of long time steps can be learned to predict future load data i 2, the two-dimensional hole causal convolutional neural network model M (·) parameters can be subtracted to learn more past values. Referring to fig. 3, a diagram of a two-dimensional hole convolutional neural network provided in an embodiment of the present invention is shown. Fig. 3 shows the structure when L is 3, each layer is convolved by two matrices into the matrix of the next layer. In fig. 3, historical load heat data of past 8 moments are used for predicting load heat at a future moment, so that the model is constructed by using the historical load heat data of the past 8 moments
Load thermal conditions, using a parametric model to predict next D t+N A deep learning model M (-) of the load value.
Step S106: testing the result of the two-dimensional cavity causal convolutional neural network model according to the test set, if the parameters in the two-dimensional cavity causal convolutional neural network model enable the target function to be the minimum on the test set, performing step S107, otherwise, adjusting the hyper-parameters of the two-dimensional cavity causal convolutional neural network model, and returning to step S105;
the objective function of the two-dimensional cavity causal convolutional neural network model M (-) is as follows:
where W, b is a parameter for predicting the network structure, γ is the weight of the regularization term, D ture For the true value, the two-dimensional cavity causal convolutional neural network model parameters W, b that minimize the objective function are found by the adam random gradient descent method.
Step S107: and predicting according to the two-dimensional cavity causal convolution neural network model, and performing inverse normalization.
Prediction value
Where N r, network M (-) is entered by input
To obtain D pre . The inverse normalization comprises: x i ＝X′ i ×X max In the formula, X max Is the maximum of all matrix elements, X i Is the i matrix element value, fig. 4 is a diagram of the predicted result at two time points provided by the embodiment of the present invention.
According to the method, the prediction of the space-time dynamic load of the charging pile is taken as a research object, the accurate prediction of the space-time dynamic load is taken as a target, and finally, a two-dimensional cavity causal convolutional neural network prediction model which can fully consider time and space is established. According to the method for predicting the space-time dynamic load of the electric automobile, the cavity factors are added to the time dimension to which the three-dimensional convolution kernel belongs, so that a two-dimensional cavity convolution layer is formed, the model can learn the information of the space dimension, then the whole network is formed by stacking the layers, the network can be guaranteed to receive long-term historical input, the model can learn the information of the time dimension, and finally the accurate prediction of the space-time dynamic load of the electric automobile is achieved.
Since the above embodiments are all described by referring to and combining with other embodiments, the same portions are provided between different embodiments, and the same and similar portions between the various embodiments in this specification may be referred to each other. And will not be described in detail herein.
It is noted that, in this specification, relational terms such as "first" and "second," and the like, are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such circuit structure, article, or apparatus. Without further limitation, the use of the phrase "comprising an … …" to define an element does not exclude the presence of additional like elements in circuit structures, articles, or devices comprising the element.
Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
The above-described embodiments of the present application do not limit the scope of the present application.
Claims (7)
1. A method for predicting space-time dynamic load of an electric vehicle is characterized by comprising the following steps:
step S101: the load data of the charging pile is preprocessed, and the preprocessing comprises the following steps: checking missing values and abnormal values in the charging pile load data; removing the missing value and the abnormal value, and filling an effective value according to a Lagrangian difference value method;
step S102: and constructing a space-time dynamic load matrix according to the charging pile load data and the charging pile position information, wherein the method comprises the following steps: step S301: establishing coordinate axes, determining the coordinates of the charging piles, and calculating the load coverage range of each charging pile; step S302: filling the load quantities of the charging piles at the first moment in the load coverage range of all the charging piles and accumulating to obtain a load matrix at the first moment
In the formula (I), the compound is shown in the specification,
is the load quantity of the point with coordinates (x, y); step S303: repeating the step S302 according to the time sequence until all the moments in the time length T contained in the electric automobile load data construct a two-dimensional load matrix, and arranging a space-time sequence D ═ D { (D) 1 ,D 2 ,...,D T },D∈R T×X×Y Wherein X, Y is a geospatial location coordinate representing a charging device;
step S103: carrying out normalization processing on the space-time dynamic load matrix to obtain a space-time dynamic load normalization matrix, wherein the normalization processing comprises the following steps: in the space-time dynamic load normalization matrix,
in the formula, X max Is the maximum of all matrix elements, X i Is the value of i matrix element, X' i The normalized matrix is processed;
step S104: dividing the space-time dynamic load normalization matrix into a training set and a test set;
step S105: training according to the training set to obtain a two-dimensional cavity causal convolution neural network model;
step S106: testing the result of the two-dimensional cavity causal convolutional neural network model according to the test set, if the parameters in the two-dimensional cavity causal convolutional neural network model enable the target function to be the minimum on the test set, performing step S107, otherwise, adjusting the hyper-parameters of the two-dimensional cavity causal convolutional neural network model, and returning to step S105;
step S107: and predicting according to the two-dimensional cavity causal convolution neural network model, and performing inverse normalization.
2. The method for predicting the spatiotemporal dynamic load of the electric vehicle as claimed in claim 1, wherein the dividing the spatiotemporal dynamic load normalization matrix into a training set and a test set comprises: 80% of the space-time dynamic load normalization matrix data are training sets, and 20% of the space-time dynamic load normalization matrix data are training sets.
3. The method for predicting spatiotemporal dynamic load of an electric vehicle according to claim 1, wherein the training according to the training set to obtain the two-dimensional cavity causal convolutional neural network model comprises:
convolving jth data of the ith layer with x, y and z positions, wherein the convolution result is as follows:
and the convolution kernel is (2 w h), the size r of the receptive field is 2 L- 1 R i Wherein x, y and z are three-dimensional charging pile positions, and d is 2 l-1 ，R i Is the size of the first dimension of the three-dimensional convolution kernel, and R i ＝2；
And stacking the convolution results in sequence to form the two-dimensional cavity causal convolution neural network model M (·).
4. The method for predicting the spatiotemporal dynamic load of the electric automobile according to claim 3, wherein the predicting according to the two-dimensional cavity causal convolutional neural network model comprises the following steps:
prediction value
Wherein N is r.
5. The method for predicting spatiotemporal dynamic load of an electric vehicle according to claim 3, wherein the objective function of the two-dimensional void causal convolutional neural network model M (-) is:
wherein W, b is a parameter for predicting the network structure, γ is the weight of the regularization term, D ture Are true values.
6. The method for predicting the spatiotemporal dynamic load of the electric vehicle according to claim 1, wherein the adjusting of the hyper-parameters of the two-dimensional cavity causal convolutional neural network model comprises:
and obtaining a two-dimensional cavity causal convolution neural network model parameter which enables the objective function to be minimum through an adam random gradient descent method.
7. The method for predicting spatiotemporal dynamic load of an electric vehicle according to claim 1, wherein the performing of the denormalization comprises: x i ＝X′ i ×X max ，
In the formula, X max Is the maximum of all matrix elements, X i Is the i matrix element value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811360331.0A CN109508830B (en) | 2018-11-15 | 2018-11-15 | Method for predicting space-time dynamic load of electric automobile |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811360331.0A CN109508830B (en) | 2018-11-15 | 2018-11-15 | Method for predicting space-time dynamic load of electric automobile |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109508830A CN109508830A (en) | 2019-03-22 |
| CN109508830B true CN109508830B (en) | 2022-09-02 |
Family
ID=65748588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811360331.0A Active CN109508830B (en) | 2018-11-15 | 2018-11-15 | Method for predicting space-time dynamic load of electric automobile |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109508830B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113853623A (en) * | 2019-05-29 | 2021-12-28 | 西门子股份公司 | Power load prediction method, device and storage medium |
| CN110956505B (en) * | 2019-12-04 | 2021-03-16 | 腾讯科技（深圳）有限公司 | Advertisement inventory estimation method and related device |
| CN111461436A (en) * | 2020-01-13 | 2020-07-28 | 长沙理工大学 | Method for predicting space-time dynamic distribution of charging load of electric automobile |
| CN112215406B (en) * | 2020-09-23 | 2024-04-16 | 国网甘肃省电力公司电力科学研究院 | Non-invasive resident electricity load decomposition method based on time convolution neural network |
| CN112508301B (en) * | 2020-12-21 | 2024-05-17 | 北京梧桐车联科技有限责任公司 | Method, device and storage medium for predicting charge load of electric vehicle |
| CN116562476B (en) * | 2023-07-12 | 2023-10-13 | 北京中电普华信息技术有限公司 | Charging load information generation method and device applied to electric automobile |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160428A (en) * | 2015-08-19 | 2015-12-16 | 天津大学 | Planning method of electric vehicle fast-charging station on expressway |
| CN106004725A (en) * | 2016-06-23 | 2016-10-12 | 海南电力技术研究院 | Vehicle terminal and orderly charging method |
| CN106849109A (en) * | 2017-03-15 | 2017-06-13 | 国网江苏省电力公司连云港供电公司 | A kind of urban distribution network load control method accessed for scale charging pile |
| CN107176051A (en) * | 2017-06-22 | 2017-09-19 | 国网天津市电力公司 | Electric automobile charging pile management system based on space-time electricity price model |
| CN107578124A (en) * | 2017-08-28 | 2018-01-12 | 国网山东省电力公司电力科学研究院 | The Short-Term Load Forecasting Method of GRU neutral nets is improved based on multilayer |
| CN108549929A (en) * | 2018-03-29 | 2018-09-18 | 河海大学 | A kind of photovoltaic power prediction technique based on deep layer convolutional neural networks |
| CN108564206A (en) * | 2018-03-27 | 2018-09-21 | 中国农业大学 | A kind of wind power forecasting method based on distributed optimization and spatial coherence |
| CN108647834A (en) * | 2018-05-24 | 2018-10-12 | 浙江工业大学 | A kind of traffic flow forecasting method based on convolutional neural networks structure |
-
2018
- 2018-11-15 CN CN201811360331.0A patent/CN109508830B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160428A (en) * | 2015-08-19 | 2015-12-16 | 天津大学 | Planning method of electric vehicle fast-charging station on expressway |
| CN106004725A (en) * | 2016-06-23 | 2016-10-12 | 海南电力技术研究院 | Vehicle terminal and orderly charging method |
| CN106849109A (en) * | 2017-03-15 | 2017-06-13 | 国网江苏省电力公司连云港供电公司 | A kind of urban distribution network load control method accessed for scale charging pile |
| CN107176051A (en) * | 2017-06-22 | 2017-09-19 | 国网天津市电力公司 | Electric automobile charging pile management system based on space-time electricity price model |
| CN107578124A (en) * | 2017-08-28 | 2018-01-12 | 国网山东省电力公司电力科学研究院 | The Short-Term Load Forecasting Method of GRU neutral nets is improved based on multilayer |
| CN108564206A (en) * | 2018-03-27 | 2018-09-21 | 中国农业大学 | A kind of wind power forecasting method based on distributed optimization and spatial coherence |
| CN108549929A (en) * | 2018-03-29 | 2018-09-18 | 河海大学 | A kind of photovoltaic power prediction technique based on deep layer convolutional neural networks |
| CN108647834A (en) * | 2018-05-24 | 2018-10-12 | 浙江工业大学 | A kind of traffic flow forecasting method based on convolutional neural networks structure |
Non-Patent Citations (3)
| Title |
|---|
| "a Spatial-Temporal model for grid impact analysis of plug-in electric vehicles";Yunfei Mu 等;《Applied Energy》;20140228;第114卷;第456-465页 * |
| "基于考虑时空分布的电动汽车充电负荷预测方法";肖建华 等;《自动化应用》;20161125(第11期);第66-67页 * |
| "考虑时空分布的电动汽车充电负荷预测方法";张洪财 等;《电力系统自动化》;20140110;第38卷(第1期);第13-20页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109508830A (en) | 2019-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109508830B (en) | Method for predicting space-time dynamic load of electric automobile | |
| Premalatha et al. | Prediction of solar radiation for solar systems by using ANN models with different back propagation algorithms | |
| CN110738252B (en) | Space autocorrelation machine learning satellite precipitation data downscaling method and system | |
| Yang et al. | Real-time profiling of fine-grained air quality index distribution using UAV sensing | |
| Millington et al. | The comparison of GEV, log-Pearson type 3 and Gumbel distributions in the Upper Thames River watershed under global climate models | |
| Barve et al. | The crucial role of the accessible area in ecological niche modeling and species distribution modeling | |
| Subedi et al. | Application of a hybrid cellular automaton–Markov (CA-Markov) model in land-use change prediction: a case study of Saddle Creek Drainage Basin, Florida | |
| US20200273347A1 (en) | Joint order dispatching and fleet management for online ride-sharing platforms | |
| JP6384065B2 (en) | Information processing apparatus, learning method, and program | |
| CN104376389B (en) | Master-slave mode microgrid power load prediction system and method based on load balancing | |
| CN110046787A (en) | A kind of urban area charging demand for electric vehicles spatio-temporal prediction method | |
| CN110400128B (en) | Spatial crowdsourcing task allocation method based on worker preference perception | |
| CN111199270A (en) | Regional wave height forecasting method and terminal based on deep learning | |
| CN107688906A (en) | The transmission line of electricity meteorological element NO emissions reduction analysis system and method for multi-method fusion | |
| DE102012219362A1 (en) | Thermal management of data centers | |
| CN105243435A (en) | Deep learning cellular automaton model-based soil moisture content prediction method | |
| CN111861028A (en) | Method for predicting crime number based on spatio-temporal data fusion | |
| CN114493052B (en) | Multi-model fusion self-adaptive new energy power prediction method and system | |
| CN115545334B (en) | Land utilization type prediction method and device, electronic equipment and storage medium | |
| CN107590570A (en) | A kind of bearing power Forecasting Methodology and system | |
| CN107358059A (en) | Short-term photovoltaic energy Forecasting Methodology and device | |
| CN114282704A (en) | Charging load prediction method and device for charging station, computer equipment and storage medium | |
| CN112508734B (en) | Method and device for predicting power generation capacity of power enterprise based on convolutional neural network | |
| CN116960962A (en) | Mid-long term area load prediction method for cross-area data fusion | |
| CN117131991A (en) | Urban rainfall prediction method and platform based on hybrid neural 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 | ||
| GR01 | Patent grant |