CN117634931B - Electric automobile adjustment capability prediction method and system considering charging behavior - Google Patents
Electric automobile adjustment capability prediction method and system considering charging behavior Download PDFInfo
- Publication number
- CN117634931B CN117634931B CN202410101586.4A CN202410101586A CN117634931B CN 117634931 B CN117634931 B CN 117634931B CN 202410101586 A CN202410101586 A CN 202410101586A CN 117634931 B CN117634931 B CN 117634931B
- Authority
- CN
- China
- Prior art keywords
- adjustment capability
- electric automobile
- data
- charge
- discharge
- 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 66
- 238000012549 training Methods 0.000 claims abstract description 149
- 230000033228 biological regulation Effects 0.000 claims abstract description 21
- 238000007599 discharging Methods 0.000 claims abstract description 17
- 230000006399 behavior Effects 0.000 claims abstract description 15
- 230000006870 function Effects 0.000 claims description 40
- 230000008569 process Effects 0.000 claims description 20
- 230000001105 regulatory effect Effects 0.000 claims description 13
- 238000012163 sequencing technique Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 10
- 239000008186 active pharmaceutical agent Substances 0.000 claims description 8
- 230000006872 improvement Effects 0.000 claims description 8
- 238000010219 correlation analysis Methods 0.000 claims description 7
- 230000004931 aggregating effect Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000005315 distribution function Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims 2
- 230000003993 interaction Effects 0.000 abstract description 7
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000001276 controlling effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000379 polymerizing effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Landscapes
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The invention discloses an electric automobile regulation capacity prediction method and system considering charging behaviors, and relates to the field of electric automobile charging regulation. An electric vehicle regulation ability prediction method considering charging behavior, comprising: acquiring charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile; collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data, charge and discharge capability characteristic data and adjustment capability of the electric automobile; training the multiple groups of adjustment capability prediction training data to obtain an adjustment capability prediction training model; inputting the data to be tested into an adjustment capability prediction training model to obtain the adjustment capability of the electric automobile. The method can excite the large-scale prediction of the adjustment capability of the electric automobile, thereby supporting the electric automobile to participate in the automobile-network interaction to realize the adjustment of the power grid.
Description
Technical Field
The invention relates to the field of electric automobile charging adjustment, in particular to an electric automobile adjustment capacity prediction method and system considering charging behaviors.
Background
The electric automobile has huge advantage as demand side regulation resource: (1) In the V2G state, the electric automobile can participate in real-time regulation and control of the power grid through charge and discharge; (2) The regulating speed of the electric automobile is far faster than the response speed of the traditional power supply; (3) The electric automobile adjusting cost is far lower than the construction and operation cost of the centralized chemical energy storage. With the great development of electric vehicles, the electric vehicle aggregate charging increases the load of a power grid, and the running flexibility of the power grid can be improved while the large-scale access of the electric vehicles is solved, and the running challenge of the urban power grid is caused by the vehicle-network interaction flexible regulation and control means. However, the participation of electric vehicles in grid regulation has the following problems: (1) The scheduling capacity of a single electric automobile is small and is insufficient for participating in the regulation and control of a power grid; (2) The travel and charging of the electric automobile user have uncertainty, and are difficult to accurately predict; (3) The method for formulating the daily plan declaration curve and the like of the electric power peak shaving auxiliary service market is rough, is formed by predicting the historical load data, and does not fully analyze and mine the electric automobile regulation capacity of regulating and controlling resources; (4) The current car-network interaction system can not meet the requirements of a multi-element interaction scene of a power grid. Therefore, there is still a need to design a method and a system for predicting the adjustment capability of an electric vehicle in consideration of charging behavior, so that the method and the system can excite the large-scale prediction of the adjustment capability of the electric vehicle, thereby supporting the electric vehicle to participate in vehicle-network interaction to realize power grid adjustment.
Disclosure of Invention
The invention aims to provide a method and a system for predicting the adjustment capability of an electric automobile, which are used for solving the technical problems by considering charging behaviors.
The invention is realized by the following technical scheme:
an electric vehicle regulation ability prediction method considering charging behavior, comprising: acquiring charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile; collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, the charge and discharge capability characteristic data and the adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model; and inputting data to be tested into the adjustment capability prediction training model to obtain the adjustment capability of the electric automobile.
The obtaining of the charging and discharging capability characteristic data having the greatest influence on the prediction of the adjustment capability of the electric automobile comprises the following steps: acquiring a plurality of charge and discharge characteristic data to be selected; analyzing the correlation between any charge and discharge characteristic data to be selected and the adjustment capacity of the electric automobile; and sequencing the plurality of charge and discharge characteristic data to be selected by utilizing the correlation analysis result, and selecting the plurality of charge and discharge characteristic data to be selected as the charge and discharge capacity characteristic data according to the sequencing.
And analyzing the correlation between any of the charge and discharge characteristic data to be selected and the adjustment capability of the electric automobile, wherein a spearman correlation coefficient is adopted, and the spearman correlation coefficient is expressed as:
；
in the method, in the process of the invention,representing the spearman correlation coefficient; />Representing the number of the charge-discharge capacity characteristic data; />Representing the charge-discharge characteristic data to be selected, < >>Representing the electric automobile regulation capability; />And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；/>And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；/>And->Respectively indicate->And->Is a rank of (2); />And->Respectively indicate->And->Average rank of (2);
the sorting the plurality of candidate charge and discharge characteristic data by using the correlation analysis result includes: and sequencing the plurality of the charge and discharge characteristic data to be selected according to the spearman correlation coefficient from large to small.
The above-mentioned collection multiunit adjustment ability prediction training data, every above-mentioned adjustment ability prediction training data of group all includes electric automobile's historical charge and discharge data, above-mentioned charge and discharge ability characteristic data and above-mentioned electric automobile adjustment ability, and every above-mentioned electric automobile adjustment ability of adjustment ability prediction training data of group obtains through following step: obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, and aggregating the multiple groups of upper and lower automobile power boundaries of each electric automobile to obtain upper and lower power boundary curves of each electric automobile; and taking the power upper and lower limit curves of various electric vehicles as the adjustment capacity of the electric vehicle of each group of the adjustment capacity prediction training data.
The training of the plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model comprises the following steps: judging whether the current input information belongs to the discarded information of the cell state of the regulation capacity prediction training model or not by the forgetting gate structure; judging whether the current input information is new information of the cell state added into the regulation capacity prediction training model or not by an input gate structure; and obtaining whether the current output information belongs to the adjustment capability of the electric automobile output by the adjustment capability prediction training model or not according to the output door structure.
The training of the plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model comprises the following steps: learning the historical charge and discharge data by adopting a Bayesian optimizer to obtain a super-parameter combination with the greatest influence on the adjustment capability of the electric automobile; and optimizing the adjustment capacity prediction training model according to the obtained hyper-parameter combination.
The historical charge and discharge data are learned by a Bayesian optimizer to obtain a super-parameter combination with the greatest improvement on the adjustment capability of the electric automobile, and the method comprises the following steps: the Bayesian optimizer learns based on historical charge and discharge data, and builds a priori function of the super parameter by adopting a Gaussian process, wherein the priori function is obtained based on the average absolute error and variance of the adjustment capability of the electric vehicle; and obtaining a posterior function by modifying the mean absolute error and the variance solution, and obtaining the super-parameter combination which minimizes the adjustment capability error of the electric automobile.
Obtaining a posterior function by modifying the mean absolute error and the variance solution, and obtaining the super-parameter combination which minimizes the adjustment capability error of the electric automobile, wherein the super-parameter combination comprises the following steps: and constructing an EI sampling function based on the average absolute error and variance of the electric automobile adjusting capability, wherein the EI sampling function is expressed as:
；
in the method, in the process of the invention,the super parameter combination which represents the minimum adjustment capability error of the electric automobile; />Representation ofThe average absolute error below; />Representing a standard normal distribution; />Representation->Probability density functions of (2); />Representation->Is a variance of (2); />Representation->Is a distribution function of (a);
the obtaining the super-parameter combination for minimizing the adjustment capability error of the electric automobile by modifying the gaussian distribution of the prior function comprises the following steps:
will beSubstituting the EI sampling function to calculate the average absolute errors of a plurality of groups of super-parameter samples; and selecting one of the super-parameter samples as the super-parameter combination with the smallest adjustment capability error of the electric automobile according to the average absolute error.
The training of the plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model comprises the following steps: inputting a plurality of groups of the adjustment capability prediction training data into an LSTM model for training to obtain the adjustment capability prediction training model; the adjustment capability prediction training model adopts an average variance loss function as a training target, and the average variance loss function is expressed as:
；
in the method, in the process of the invention,representing an average variance loss value; />Representing a plurality of sets of the adjustment capability prediction training data;indicate->The adjustment capability of the real electric automobile; />Indicate->The electric automobile can adjust the capacity.
An electric vehicle regulation capability prediction system taking charging behavior into account, comprising: and a feature screening module: the method comprises the steps of acquiring characteristic data of charge and discharge capacity with the largest influence on the prediction of the adjustment capacity of the electric automobile; model training module: the device comprises a data acquisition module, a data acquisition module and a data processing module, wherein the data acquisition module is used for acquiring a plurality of groups of adjustment capability prediction training data, and each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, the charge and discharge capability characteristic data and the adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model; capability prediction module: and the data to be tested is input into the adjustment capability prediction training model to obtain the adjustment capability of the electric automobile.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the method, the characteristic data of the charge and discharge capacity with the largest influence on the prediction of the adjustment capacity of the electric automobile are obtained; collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data, charge and discharge capability characteristic data and adjustment capability of the electric automobile; training the multiple groups of adjustment capability prediction training data to obtain an adjustment capability prediction training model; and obtaining the electric automobile adjustment capability of the data to be tested through the adjustment capability prediction training model. The charging and discharging capability characteristic data which has the greatest influence on the electric vehicle adjusting capability is utilized for learning, a model for predicting the electric vehicle adjusting capability is obtained, and the large-scale prediction of the electric vehicle adjusting capability is stimulated, so that the electric vehicle is supported to participate in vehicle-network interaction to realize power grid adjustment.
Drawings
In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the drawings that are needed in the examples will be briefly described below, it being understood that the following drawings only illustrate some examples of the present invention and therefore should not be considered as limiting the scope, and that other related drawings may be obtained from these drawings without inventive effort for a person skilled in the art. In the drawings:
FIG. 1 is a schematic flow chart of the Bayesian optimizer of the embodiment 1 of the present application;
FIG. 2 is a graph showing the upper and lower power curves of example 1 of the present application;
fig. 3 is a schematic diagram of LSTM neural network in embodiment 1 of the present application.
Detailed Description
For the purpose of making apparent the objects, technical solutions and advantages of the present invention, the present invention will be further described in detail with reference to the following examples and the accompanying drawings, wherein the exemplary embodiments of the present invention and the descriptions thereof are for illustrating the present invention only and are not to be construed as limiting the present invention.
Example 1
As shown in fig. 1 to 3, an embodiment of the present application provides a method for predicting adjustment capability of an electric vehicle in consideration of charging behavior, including: acquiring charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile; collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, the charge and discharge capability characteristic data and the adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model; and inputting data to be tested into the adjustment capability prediction training model to obtain the adjustment capability of the electric automobile.
The electric automobile adjusting capability is defined as the size of the electric automobile capable of providing flexible up-adjusting and down-adjusting capability when a power grid has certain adjusting requirements. Whether the electric automobile can participate in power grid adjustment after receiving the scheduling signal is related to the electric quantity, the charging and discharging power and the charging stopping time of the electric automobile at the moment.
The obtaining of the charging and discharging capability characteristic data having the greatest influence on the prediction of the adjustment capability of the electric automobile comprises the following steps: acquiring a plurality of charge and discharge characteristic data to be selected; analyzing the correlation between any charge and discharge characteristic data to be selected and the adjustment capacity of the electric automobile; and sequencing the plurality of charge and discharge characteristic data to be selected by utilizing the correlation analysis result, and selecting the plurality of charge and discharge characteristic data to be selected as the charge and discharge capacity characteristic data according to the sequencing.
Optionally, when acquiring the plurality of charge and discharge feature data to be selected by collecting user order data of the electric vehicle charging station, the charge and discharge feature data to be selected may include charge related factors: such as charge start-stop time, charge start-stop SOC, electric vehicle battery capacity, and maximum charge rate of the electric vehicle. The data can be preprocessed by utilizing the factors, and abnormal data in the order data, such as data with zero charge quantity and charge duration, can be deleted.
And re-sequencing and encoding the data to obtain data which can be identified by the program, thereby extracting and obtaining a plurality of charge and discharge characteristic data to be selected. For example, the data may be normalized using the mean variance.
Optionally, the data is normalized by using the mean variance to obtain historical charge and discharge data of each group of adjustment capability prediction training data, which is expressed as:
；
in the method, in the process of the invention,for the value before normalization, ++>For the normalized value, μ is the average value of the historical charge and discharge data of the plurality of sets of adjustment capability prediction training data, and S is the standard deviation of the historical charge and discharge data of the plurality of sets of adjustment capability prediction training data.
After normalization processing is carried out on the acquired data, the data can be divided into a plurality of groups according to different time periods; and each group of the adjustment capability prediction training data is used for acquiring historical charge and discharge data and multiple charge and discharge capability characteristic data of each electric automobile through each group of data, and calculating according to the next group of data to obtain the electric automobile adjustment capability of each electric automobile. Alternatively, the multiple sets of adjustment capability prediction training data may be divided into a training set and a validation set of adjustment capability prediction training models.
And analyzing the correlation between any of the charge and discharge characteristic data to be selected and the adjustment capability of the electric automobile, wherein a spearman correlation coefficient is adopted, and the spearman correlation coefficient is expressed as:
；
in the method, in the process of the invention,representing the spearman correlation coefficient; />Representing the number of the charge-discharge capacity characteristic data; />Representing the charge-discharge characteristic data to be selected, < >>Representing the electric automobile regulation capability; />And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；/>And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；/>And->Respectively indicate->And->Is a rank of (2); />And->Respectively indicate->And->Average rank of (c).
The sorting the plurality of candidate charge and discharge characteristic data by using the correlation analysis result includes: and sequencing the plurality of the charge and discharge characteristic data to be selected according to the spearman correlation coefficient from large to small.
Optionally, before analyzing the correlation between the plurality of factors and the flexible adjustment capability of the electric automobile and performing feature selection, (1) collecting factor data related to the flexible adjustment capability of the electric automobile. The charging state information of the electric vehicle, such as charging start-stop time, charging start-stop SOC, electric vehicle battery capacity, and maximum charging rate of the electric vehicle, and charging condition information may be included. The charging condition information includes the number of electric vehicles charged at a certain moment, the number of electric vehicles starting to be charged at a certain moment, and the like; weather information (cloudy, sunny, rainy), temperature, wind speed and the like related to weather forecast; time-dependent weekdays and weekends, seasons, months, dates and times, etc. (2) then performing data characterization processing: for the time feature, the time feature can be disassembled to be converted into discrete time features such as year, month, day, time, minute, day of week, etc. The class type features, such as weather, and the numerical type features, such as the charging quantity of the electric automobile at a certain moment, can be converted into the numerical type features by adopting an encoding mode of the Embedding, and are used as prediction spiritInput characteristics via the network. The Embedding network can convert the large sparse vector into a low-dimensional vector which keeps the relation among the features, can capture the inherent attribute of the original features and discover the relation among the vectors, and can further improve the accuracy of flexible adjustment capability prediction by using the feature vector processed by the Embedding network. And then performing (3) feature selection based on the processed data, namely: the most representative feature is selected from the existing features and the useless or redundant features are removed. By screening and retaining the most important features, the interference of noise or irrelevant features can be eliminated, and the prediction accuracy and precision of the model are improved. Meanwhile, when the number of the features is large, the problem of over fitting of the model is easy to occur, the number of the features can be reduced through feature selection, the complexity of the model is reduced, and therefore the risk of over fitting is reduced. In addition, feature selection has the marry of reducing the amount of data that needs to be processed, thereby shortening training time and saving computing resources. The correlation of the existing characteristics and the flexible adjustment capability is analyzed by adopting the spearman correlation coefficient, the statistic for measuring the tightness degree of the linear relationship between the two variables can be set between-1 and 1, wherein 1 represents complete positive correlation, -1 represents complete negative correlation, and 0 represents no correlation. The spearman correlation examines the strength of the monotonic relationship between the two, namely how the two keep pace in the trend of increasing or decreasing colloquially, and the spearman correlation coefficient is calculated by using the definition of the ranking order value of the data sampleAnd->Two sets of data. Alternatively, the present invention may also determine the final best feature combination based on a comparison of different feature data subsets, overall evaluation. And constructing a corresponding neural network by utilizing important characteristics which can influence the flexible adjustment capability of the electric automobile most, so as to realize the prediction of the flexible adjustment capability of the electric automobile.
The above-mentioned collection multiunit adjustment ability prediction training data, every above-mentioned adjustment ability prediction training data of group all includes electric automobile's historical charge and discharge data, above-mentioned charge and discharge ability characteristic data and above-mentioned electric automobile adjustment ability, and every above-mentioned electric automobile adjustment ability of adjustment ability prediction training data of group obtains through following step: obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, and aggregating the multiple groups of upper and lower automobile power boundaries of each electric automobile to obtain upper and lower power boundary curves of each electric automobile; and taking the power upper and lower limit curves of various electric vehicles as the adjustment capacity of the electric vehicle of each group of the adjustment capacity prediction training data.
Alternatively, to accurately represent the adjustment capability of the electric vehicle, the following aspects may be described. (1) And according to the charging rule of the electric automobile, formulating a charging mode under the condition that the electric automobile does not participate in regulation and control. For example, if there is no external interference, the default electric automobile user charges with maximum power after connecting the charging pile until reaching the required electric quantity. (2) The response direction and duration of the electric vehicle demand response signal can be given, and the upper and lower power limits of the electric vehicle can be determined in combination with the electric vehicle charging state. For example, an electric vehicle participates in demand response on the premise that its energy and power are constrained by:
；
；
；
；
；
the energy upper and lower limits of each electric automobileMaximum energy of the battery of the electric vehicle is applied +.>And minimum energy->Is a constraint of (2); power upper and lower bound->Maximum charging power of the battery of the electric vehicle>And maximum discharge power +.>Is a constraint of (a). Wherein, the maximum energy of the battery of the electric automobile is +.>The battery capacity and the minimum value of the electric quantity of the electric vehicle, which is reached by the maximum power charging energy, are the minimum energy of the battery of the electric vehicle +.>And the power upper and lower bounds of the electric automobile are obtained according to the discharging capacity of the battery and the relation between the battery storage energy and the electric automobile power, so that the travel requirement of the electric automobile is met.
(3) And polymerizing the upper and lower power boundaries of different types of electric automobile clusters to obtain upper and lower power boundary curves of various electric automobiles. And when DS= -1, the electric automobile is required to discharge or reduce the charging power, meanwhile, the duration of the regulating signal is defined, and the constraint on the power energy of the electric automobile is regulated according to the power grid signal, so that the upper power limit and the lower power limit of the electric automobile are obtained. See in particular FIG. 2 below for eachThe Power upper and lower boundary curve diagram of the electric automobile is shown, DS represents a Power grid regulating signal, power represents energy, and SOC represents the battery capacity of the electric automobile. In the drawing the view of the figure,is the start charging time, +.>Ending the charging time, and selecting the electric vehicle to charge with the maximum power when no power grid adjusting signal exists, until the electric vehicle reaches the electric quantity requirement; when ds=1, if the electric quantity requirement and the travel requirement of the electric automobile are met, the electric automobile can translate the charging time, and charge with the maximum power in the period of ds=1; when ds= -1, if the electric power demand and travel demand of the electric vehicle are satisfied, the electric vehicle can be discharged and not charged in this period. Therefore, the electric vehicle power is obtained when the upper power limit is ds=1, and the electric vehicle power is obtained when the lower power limit is ds= -1, and the electric vehicle power is positive when charging and negative when discharging.
After the upper and lower power limits of each electric automobile are obtained, the adjustment capability of a plurality of electric automobiles of the same type is summed to obtain the upper and lower power limits of the electric automobiles, namely the adjustment capability (potential) of the electric automobiles, and the following formula is specifically referenced:
；
in the method, in the process of the invention,indicate->The same electric automobile corresponding to the electric automobile adjusting capability>Power upper and lower bound of->Indicating the number of such electric vehicles.
In order to improve the prediction accuracy, the prediction result can be obtained for a plurality of times based on the adjustment capability prediction training model, and the probability prediction of the flexible adjustment capability of the electric automobile is respectively carried out at the confidence levels of 80%, 90% and 95%. And, the prediction result can be evaluated by selecting an index of the average bandwidth (PINAW) of the prediction interval, which is expressed as:
；
in the method, in the process of the invention,represents the extreme difference of the real adjustment capability, K represents the group number of the predictive training data of the multi-group adjustment capability,/and->And->Respectively represent->The smaller the PINAW, the better the prediction effect is represented by the upper power limit and the lower power limit corresponding to the adjustment capability interval of the electric automobile.
As shown in fig. 3, when the LSTM neural network is applied to obtain the capacity adjustment prediction training model, the training of the plurality of sets of the capacity adjustment prediction training data to obtain the capacity adjustment prediction training model includes: judging whether the current input information belongs to the discarded information of the cell state of the regulation capacity prediction training model or not by the forgetting gate structure; judging whether the current input information is new information of the cell state added into the regulation capacity prediction training model or not by an input gate structure; and obtaining whether the current output information belongs to the adjustment capability of the electric automobile output by the adjustment capability prediction training model or not according to the output door structure.
The discard information of the cell state of the regulatory capability prediction training model is obtained by forgetting the gate structure and is expressed as follows:
；
the gate is a method of selectively passing information and includes a sigmoid neural network layer and a nonlinear operation of a pointwise multiplication. In the aboveThe current input information representing the adjustment capability prediction training model; />Representation->Mapping to the degree of forgetting and retention; />The last output information of the predictive training model for regulating ability; />A nonlinear mapping representing a Sigmoid function, the value of which is {0,1}; />Representation->Weights of the training model are predicted on the above regulatory capacity, +.>Representation->Deviation of the adjustment capability prediction training model; wherein, when->When 0, the drug is added>Belonging to the above-mentioned discarding information; when->1->Not belonging to the discard information.
New information of the cell state of the above regulatory capability prediction training model is obtained from the input gate structure, expressed as:
；
；
in the method, in the process of the invention,indicating a determination as to whether the current input information requires updating of the cell state,representation ofThe above new information on whether or not it belongs to the cell state;、representing the weight of the corresponding current input information on the adjustment capability prediction training model,、representing the deviation of the corresponding current input information from the adjustment capability prediction training model; when (when)Andand if the result is positive, updating the cell state. Among them, tanh also activates functions, mainly selecting cell states, leaving unnecessary, and remaining useful.
The electric automobile adjusting capacity output by the adjusting capacity prediction training model is obtained through an output door structure and is expressed as:
；
；
in the method, in the process of the invention,indicating whether the current electric automobile regulating capability is output or not; />The output adjustment capability of the electric automobile is represented; />Weight of the predictive training model representing the corresponding current input information on the above-mentioned adjustment capability, +.>Representing the deviation of the corresponding current input information from the adjustment capability prediction training model.
The training of the plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model comprises the following steps: learning the historical charge and discharge data by adopting a Bayesian optimizer to obtain a super-parameter combination with the maximum improvement of the adjustment capability of the electric automobile; and optimizing the adjustment capacity prediction training model according to the obtained hyper-parameter combination.
The learning of the historical charge and discharge data by using a bayesian optimizer to obtain a super-parameter combination with the greatest improvement on the adjustment capability of the electric automobile comprises the following steps: the Bayesian optimizer learns based on the historical charge and discharge data, and builds a priori function of the super parameter by adopting a Gaussian process, wherein the priori function is obtained based on the average absolute error and variance of the adjustment capability of the electric automobile; and solving the prior function by modifying the average absolute error and the variance to obtain the super-parameter combination which minimizes the adjustment capability error of the electric automobile. Wherein the prior function is expressed as:
；
in the method, in the process of the invention,representing the above-mentioned super parameter combinations; />Representing true regulatory capacity and +.>Average absolute error of the adjustment capability of the electric automobile; />Representation->The minimum value of the average absolute error of the real adjustment capability and the electric vehicle adjustment capability; />A covariance function representing the adjustment capability of the electric vehicle; and solving the prior function by modifying the average absolute error and the variance to obtain the super-parameter combination which minimizes the adjustment capability error of the electric automobile.
Solving the prior function by modifying the average absolute error and the variance to obtain the super-parameter combination which minimizes the adjustment capability error of the electric automobile, comprising: and constructing an EI sampling function based on the average absolute error and variance of the electric automobile adjusting capability, wherein the EI sampling function is expressed as:
；
in the method, in the process of the invention,the super parameter combination which represents the minimum adjustment capability error of the electric automobile; />Representation ofThe average absolute error below; />Representing a standard normal distribution; />Representation->Probability density functions of (2); />Representation->Is a variance of (2); />Representation->Is a distribution function of (a);
the obtaining the super-parameter combination for minimizing the adjustment capability error of the electric automobile by modifying the gaussian distribution of the prior function comprises the following steps:
will beSubstituting the EI sampling function to calculate the average absolute errors of a plurality of groups of super-parameter samples; and selecting one of the super-parameter samples as the super-parameter combination with the smallest adjustment capability error of the electric automobile according to the average absolute error.
The flow of the Bayesian optimizer is shown in fig. 1, and the Bayesian optimizer finds the neural network hyper-parameter combination which minimizes the objective function value by constructing a proxy model (a priori function). Taking a Gaussian process as a proxy model, and continuously updating the prior by considering the previous model parameter information; and learning by adopting historical charge and discharge data, namely power upper and lower bounds of the electric automobile cluster, and finding out the model parameter with the maximum improvement of the prediction result to optimize. The design idea of the EI sampling function is to find the parameter combination expected by the maximum improvement of the next prediction result, and calculate the combination by a sampleConverting into standard normal distribution, and finding out optimal local solution to obtain the super-parameter combination ++when the adjustment capability error of the electric automobile is minimum>。
The training of the plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model comprises the following steps: inputting a plurality of groups of the adjustment capability prediction training data into an LSTM model for training to obtain the adjustment capability prediction training model; the adjustment capability prediction training model adopts an average variance loss function as a training target, and the average variance loss function is expressed as:
；
in the method, in the process of the invention,representing an average variance loss value; />Representing a plurality of sets of the adjustment capability prediction training data;indicate->The adjustment capability of the real electric automobile; />Indicate->The electric automobile can adjust the capacity.
Since the parameter setting of LSTM may cause a problem of poor network fitting when only using LSTM to predict flexible adjustability, a bayesian optimizer is introduced to optimize super parameters such as learning rate, number of hidden layer nodes, and regularization factor of LSTM network. The network randomness caused by the empirical determination of the network parameters is avoided, and the prediction accuracy of the network is relatively improved.
Example 2
The embodiment of the application provides an electric automobile adjustment capability prediction system considering charging behavior, which comprises: and a feature screening module: the method comprises the steps of acquiring characteristic data of charge and discharge capacity with the largest influence on the prediction of the adjustment capacity of the electric automobile; model training module: the device comprises a data acquisition module, a data acquisition module and a data processing module, wherein the data acquisition module is used for acquiring a plurality of groups of adjustment capability prediction training data, and each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, the charge and discharge capability characteristic data and the adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model; capability prediction module: and the data to be tested is input into the adjustment capability prediction training model to obtain the adjustment capability of the electric automobile.
The principle of the embodiment of the present application is the same as that of embodiment 1, and a repetitive description thereof will not be made here.
In summary, the embodiment of the application provides a method and a system for predicting the adjustment capability of an electric vehicle by considering charging behaviors: acquiring charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile; collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data, charge and discharge capability characteristic data and adjustment capability of the electric automobile; training the multiple groups of adjustment capability prediction training data to obtain an adjustment capability prediction training model; and obtaining the electric automobile adjustment capability of the data to be tested through the adjustment capability prediction training model. The charging and discharging capability characteristic data with the largest influence on the electric vehicle adjusting capability is utilized for learning, a model for predicting the electric vehicle adjusting capability is obtained, the large-scale prediction of the electric vehicle adjusting capability is stimulated, and the electric vehicle is supported to participate in interaction between the electric vehicle and a power grid, so that the power grid adjusting and controlling is realized.
According to the embodiment of the invention, the historical charge and discharge data of the electric automobile are utilized, the relation between power and energy is considered, the electric automobile adjusting capacity of the electric automobile is depicted, and the electric automobile adjusting capacity of the large-scale electric automobile is aggregated; considering factors influencing the adjustment capacity of the electric automobile, and analyzing the adjustment capacity of the electric automobile by using important characteristics; and carrying out probability prediction on the electric automobile adjusting capacity by adopting a Bayesian optimized LSTM model to obtain the electric automobile adjusting capacity under different confidence intervals, and fully releasing flexible resources of the power distribution network.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (8)
1. The electric automobile regulation capacity prediction method considering charging behavior is characterized by comprising the following steps of:
acquiring charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile;
collecting multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, charging and discharge capability characteristic data and adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model;
inputting data to be tested into the adjustment capability prediction training model to obtain the adjustment capability of the electric automobile;
the method comprises the steps that a plurality of groups of adjustment capability prediction training data are collected, each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, charge and discharge capability characteristic data and electric automobile adjustment capability, and the electric automobile adjustment capability of each group of adjustment capability prediction training data is obtained through the following steps:
obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, and aggregating multiple groups of upper and lower automobile power boundaries of different types of electric automobile clusters to obtain power upper and lower boundary curves of various electric automobiles;
taking the power upper and lower limit curves of various electric vehicles as the electric vehicle adjusting capacity of each group of the adjusting capacity prediction training data;
obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, setting a power grid adjusting signal as DS, increasing charging power of the electric automobiles when DS=1, discharging or reducing the charging power of the electric automobiles when DS= -1, and defining duration time of the adjusting signal to obtain an upper power limit and a lower power limit of the electric automobiles;
the obtaining of the charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile comprises the following steps:
acquiring a plurality of charge and discharge characteristic data to be selected;
analyzing the correlation between any charge and discharge characteristic data to be selected and the adjustment capacity of the electric automobile;
and sequencing the plurality of charge and discharge characteristic data to be selected by utilizing a correlation analysis result, and selecting the plurality of charge and discharge characteristic data to be selected as the charge and discharge capacity characteristic data according to sequencing.
2. The method for predicting the adjustment capability of the electric vehicle according to claim 1, wherein the analyzing the correlation between any of the candidate charge-discharge characteristic data and the adjustment capability of the electric vehicle uses a spearman correlation coefficient, and the spearman correlation coefficient is expressed as:
；
in the method, in the process of the invention,representing the spearman correlation coefficient; />Representing the number of the charge-discharge capacity characteristic data; />Representing the charge-discharge characteristic data to be selected, < >>Representing the electric automobile regulation capability; />And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；/>And->Respectively represent +.>The corresponding adjusting capacity of the electric automobile is +.>And->；And->Respectively indicate->And->Is a rank of (2); />And->Respectively indicate->And->Average rank of (2);
the sorting the plurality of the charge and discharge characteristic data to be selected by using the correlation analysis result comprises the following steps:
and sequencing the plurality of charge and discharge characteristic data to be selected according to the spearman correlation coefficient from large to small.
3. The method for predicting the adjustment capability of the electric vehicle according to claim 1, wherein the plurality of sets of adjustment capability prediction training data are trained to obtain an adjustment capability prediction training model, and an LSTM algorithm is adopted, and the method comprises:
judging whether the current input information belongs to the discarded information of the cell state of the regulatory capability prediction training model or not by a forgetting door structure;
judging whether the current input information is new information of the cell state added into the regulatory capability prediction training model or not by an input gate structure;
and obtaining whether the current output information belongs to the electric automobile adjusting capacity output by the adjusting capacity prediction training model or not according to the output door structure.
4. The method for predicting the adjustment capability of an electric vehicle according to claim 3, wherein the plurality of sets of adjustment capability prediction training data are trained to obtain an adjustment capability prediction training model, comprising:
learning the historical charge and discharge data by adopting a Bayesian optimizer to obtain a super-parameter combination with the maximum improvement of the adjustment capability of the electric automobile; and optimizing the adjustment capability prediction training model according to the obtained hyper-parameter combination.
5. The method for predicting the adjustment capability of an electric vehicle with consideration of charging behavior according to claim 4, wherein the learning the historical charging and discharging data by using a bayesian optimizer to obtain a super-parameter combination with the greatest improvement of the adjustment capability of the electric vehicle comprises:
the Bayesian optimizer learns based on the historical charge and discharge data, and builds a priori function of the super-parameters by adopting a Gaussian process; and obtaining the super-parameter combination which minimizes the adjustment capability error of the electric automobile by modifying the Gaussian distribution of the prior function.
6. The method for predicting the adjustment capability of an electric vehicle with consideration of charging behavior according to claim 5, wherein said constructing an a priori function of the super-parameters using a gaussian process comprises: obtaining the super-parameter combination with the minimum adjustment capability error of the electric automobile through sample calculation, wherein the sample calculation adopts an EI sampling function, and the EI sampling function is expressed as:
；
in the method, in the process of the invention,the super-parameter combination representing the time when the adjustment capability error of the electric automobile is minimum; />Representation->The average absolute error below; />Representing a standard normal distribution; />Representation->Probability density functions of (2); />Representation->Is a variance of (2);representation->Is a distribution function of (a);
the step of obtaining the super-parameter combination which minimizes the adjustment capability error of the electric automobile by modifying the Gaussian distribution of the prior function comprises the following steps:
will beSubstituting the EI sampling function, and calculating the average absolute errors of a plurality of groups of super-parameter samples; and selecting one of the super-parameter samples as the super-parameter combination with the minimum adjustment capability error of the electric automobile according to the average absolute error.
7. The method for predicting the adjustment capability of an electric vehicle according to claim 1, wherein the plurality of sets of adjustment capability prediction training data are trained to obtain an adjustment capability prediction training model, comprising:
inputting a plurality of groups of the adjustment capability prediction training data into an LSTM model for training to obtain an adjustment capability prediction training model; the adjustment capability prediction training model adopts an average variance loss function as a training target, wherein the average variance loss function is expressed as:
；
in the method, in the process of the invention,representing an average variance loss value; />Representing a number of sets of said adjustment capability prediction training data; />Indicate->The adjustment capability of the real electric automobile; />Indicate->And the electric automobile can adjust the capacity.
8. A system for an electric vehicle regulation ability prediction method taking into account charging behavior as defined in claim 1, comprising:
and a feature screening module: the method comprises the steps of acquiring characteristic data of charge and discharge capacity with the largest influence on the prediction of the adjustment capacity of the electric automobile;
model training module: the method comprises the steps of acquiring multiple groups of adjustment capability prediction training data, wherein each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, charging and discharge capability characteristic data and adjustment capability of the electric automobile; training a plurality of groups of the adjustment capability prediction training data to obtain an adjustment capability prediction training model;
capability prediction module: the electric vehicle adjusting capacity is obtained by inputting data to be tested into the adjusting capacity prediction training model;
the method comprises the steps that a plurality of groups of adjustment capability prediction training data are collected, each group of adjustment capability prediction training data comprises historical charge and discharge data of an electric automobile, charge and discharge capability characteristic data and electric automobile adjustment capability, and the electric automobile adjustment capability of each group of adjustment capability prediction training data is obtained through the following steps:
obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, and aggregating multiple groups of upper and lower automobile power boundaries of different types of electric automobile clusters to obtain power upper and lower boundary curves of various electric automobiles;
taking the power upper and lower limit curves of various electric vehicles as the electric vehicle adjusting capacity of each group of the adjusting capacity prediction training data;
obtaining multiple groups of upper and lower automobile power boundaries of multiple electric automobiles according to multiple groups of historical charge and discharge data, setting a power grid adjusting signal as DS, increasing charging power of the electric automobiles when DS=1, discharging or reducing the charging power of the electric automobiles when DS= -1, and defining duration time of the adjusting signal to obtain an upper power limit and a lower power limit of the electric automobiles;
the obtaining of the charging and discharging capability characteristic data with the largest influence on the prediction of the adjustment capability of the electric automobile comprises the following steps:
acquiring a plurality of charge and discharge characteristic data to be selected;
analyzing the correlation between any charge and discharge characteristic data to be selected and the adjustment capacity of the electric automobile;
and sequencing the plurality of charge and discharge characteristic data to be selected by utilizing a correlation analysis result, and selecting the plurality of charge and discharge characteristic data to be selected as the charge and discharge capacity characteristic data according to sequencing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410101586.4A CN117634931B (en) | 2024-01-25 | 2024-01-25 | Electric automobile adjustment capability prediction method and system considering charging behavior |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410101586.4A CN117634931B (en) | 2024-01-25 | 2024-01-25 | Electric automobile adjustment capability prediction method and system considering charging behavior |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117634931A CN117634931A (en) | 2024-03-01 |
| CN117634931B true CN117634931B (en) | 2024-04-16 |
Family
ID=90027244
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410101586.4A Active CN117634931B (en) | 2024-01-25 | 2024-01-25 | Electric automobile adjustment capability prediction method and system considering charging behavior |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117634931B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899667A (en) * | 2015-06-30 | 2015-09-09 | 华北电力大学（保定） | Electric car charging-discharging behavior prediction method |
| CN106203720A (en) * | 2016-07-15 | 2016-12-07 | 合肥工业大学 | A kind of Multiple Time Scales electric automobile cluster schedulable capacity prediction methods |
| CN109508826A (en) * | 2018-11-13 | 2019-03-22 | 合肥工业大学 | The schedulable capacity prediction methods of electric car cluster of decision tree are promoted based on gradient |
| CN112332433A (en) * | 2020-09-29 | 2021-02-05 | 国网电动汽车服务有限公司 | Transferable load capacity analysis method for electric vehicle participated in valley filling auxiliary service |
| WO2023010759A1 (en) * | 2021-08-04 | 2023-02-09 | 国网江苏省电力有限公司南通供电分公司 | Power distribution and sale competitive situation-based regional power distribution network gridding load interval prediction method |
| CN116436080A (en) * | 2023-04-12 | 2023-07-14 | 国网湖南省电力有限公司 | Comprehensive evaluation method and system for participating in power grid regulation and control by flexible resource clusters |
| CN117096868A (en) * | 2023-08-28 | 2023-11-21 | 华北电力大学（保定） | Micro-grid energy scheduling method considering various flexible loads and electric vehicles |
| CN117291304A (en) * | 2023-09-28 | 2023-12-26 | 清华大学 | Method, device, equipment and medium for predicting charge-discharge adjustable capacity of electric automobile |
-
2024
- 2024-01-25 CN CN202410101586.4A patent/CN117634931B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899667A (en) * | 2015-06-30 | 2015-09-09 | 华北电力大学（保定） | Electric car charging-discharging behavior prediction method |
| CN106203720A (en) * | 2016-07-15 | 2016-12-07 | 合肥工业大学 | A kind of Multiple Time Scales electric automobile cluster schedulable capacity prediction methods |
| CN109508826A (en) * | 2018-11-13 | 2019-03-22 | 合肥工业大学 | The schedulable capacity prediction methods of electric car cluster of decision tree are promoted based on gradient |
| CN112332433A (en) * | 2020-09-29 | 2021-02-05 | 国网电动汽车服务有限公司 | Transferable load capacity analysis method for electric vehicle participated in valley filling auxiliary service |
| WO2023010759A1 (en) * | 2021-08-04 | 2023-02-09 | 国网江苏省电力有限公司南通供电分公司 | Power distribution and sale competitive situation-based regional power distribution network gridding load interval prediction method |
| CN116436080A (en) * | 2023-04-12 | 2023-07-14 | 国网湖南省电力有限公司 | Comprehensive evaluation method and system for participating in power grid regulation and control by flexible resource clusters |
| CN117096868A (en) * | 2023-08-28 | 2023-11-21 | 华北电力大学（保定） | Micro-grid energy scheduling method considering various flexible loads and electric vehicles |
| CN117291304A (en) * | 2023-09-28 | 2023-12-26 | 清华大学 | Method, device, equipment and medium for predicting charge-discharge adjustable capacity of electric automobile |
Non-Patent Citations (5)
| Title |
|---|
| Steven S. Skiena著.《大数据分析：理论、方法及应用》.北京机械工业出版社,2022,第29-32页. * |
| 周爱民 等.《秘境寻优：人工智能中的搜索方法》.上海科学技术文献出版社,2022,(第1版),第168-173页. * |
| 电动汽车参与区域电网需求响应优化管理研究;杨晓龙;《中国博士学位论文全文数据库 工程科技II辑》;20230315(第03期);第37-54页 * |
| 电动汽车虚拟储能在电网中的应用及商业模式研究;梁登香;《中国优秀硕士学位论文全文数据库 工程科技II辑》;20220115(第01期);第5-10页 * |
| 高志强 等.《深度学习：从入门到实践》.中国铁道出版社,2018,(第1版),第123-128页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117634931A (en) | 2024-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107766986A (en) | Leak integral form echo state network on-line study photovoltaic power Forecasting Methodology | |
| CN113554466B (en) | Short-term electricity consumption prediction model construction method, prediction method and device | |
| CN112434848A (en) | Nonlinear weighted combination wind power prediction method based on deep belief network | |
| CN111008726B (en) | Class picture conversion method in power load prediction | |
| CN112305441B (en) | Power battery health state assessment method under integrated clustering | |
| CN114511132A (en) | Photovoltaic output short-term prediction method and prediction system | |
| CN113988426A (en) | Electric vehicle charging load prediction method and system based on FCM clustering and LSTM | |
| CN115099511A (en) | Photovoltaic power probability estimation method and system based on optimized copula | |
| CN116106761A (en) | Real-time lithium ion battery electric quantity estimation method based on typical correlation analysis | |
| CN112348236B (en) | Abnormal daily load demand prediction system and method for intelligent electricity consumption monitoring terminal | |
| US20230305073A1 (en) | Method and apparatus for providing a predicted aging state of a device battery based on a predicted usage pattern | |
| CN117634931B (en) | Electric automobile adjustment capability prediction method and system considering charging behavior | |
| CN116819372A (en) | Method for calculating health state of energy storage battery by neural network | |
| CN116826710A (en) | Peak clipping strategy recommendation method and device based on load prediction and storage medium | |
| CN117034762A (en) | Composite model lithium battery life prediction method based on multi-algorithm weighted sum | |
| CN116187506A (en) | Short-term wind power combination probability prediction method and system considering meteorological classification | |
| CN115329277A (en) | SOH prediction method for movement-withdrawing nickel-metal hydride battery | |
| Tao et al. | A novel method of SOC estimation for electric vehicle based on adaptive particle filter | |
| CN113393035B (en) | Daily charge and discharge power prediction method for electric automobile | |
| CN115764855A (en) | Real-time adjustable capacity and available electric quantity prediction method for electric vehicle quick charging station | |
| CN116316542A (en) | Regional distributed photovoltaic power prediction method, regional distributed photovoltaic power prediction device, computer equipment and storage medium | |
| CN113762591B (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM countermeasure learning | |
| CN112949938B (en) | Wind power climbing event direct forecasting method for improving training sample class imbalance | |
| CN112001564B (en) | User leaving time prediction method based on long-short term memory network | |
| Li et al. | Research on the prediction method of power battery SOC based on deep learning |
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 |