CN-121615519-B - Electric vehicle charging strategy construction method and system considering user charging behavior

CN121615519BCN 121615519 BCN121615519 BCN 121615519BCN-121615519-B

Abstract

The invention discloses an electric vehicle charging strategy construction method and system considering user charging behaviors, and relates to the technical field of electric vehicle charging. The method comprises the steps of collecting original multidimensional data comprising charging load, travel time, weather and the like, restoring the charging load data through parallel factor decomposition of a structured tensor, predicting travel time through a xLSTM model after dimension reduction, and outputting optimal charging power distribution and charging decisions of each time step by adopting a two-layer optimization strategy combining a sparrow search algorithm based on complete load data and travel time prediction results. According to the invention, the power grid overload is avoided through multi-dimensional constraint, the user satisfaction is taken as an optimization target, the number of vehicles which do not reach the target SOC is obviously reduced, the user charging experience is improved, and the stable operation of the power grid is ensured.

Inventors

LI CHUN
WANG CHAO
ZHANG ZHIDA
Shang Shuonan
WANG JINGPENG
CHEN XI
LI CHAOYANG
ZHANG XIAOWEN
HE JINZHAO
NIU SHUYA

Assignees

国网天津市电力公司城西供电分公司
国网天津市电力公司
国家电网有限公司

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (9)

1. The electric automobile charging strategy construction method considering the charging behavior of the user is characterized by comprising the following steps: Step1, collecting original multidimensional data; The original multidimensional data comprise electric vehicle charging load data, electric vehicle arrival time historical observation data, electric vehicle departure time historical observation data, meteorological data, battery charge state data and target SOC; Step 2, repairing the electric vehicle charging load data in the original multidimensional data by using an electric vehicle load data repairing model based on structured tensor parallel factor decomposition to obtain complete electric vehicle charging load data; step 3, performing dimension reduction processing on the original multidimensional data to obtain dimension reduced data, and predicting by using an electric vehicle travel time prediction model based on xLSTM based on the dimension reduced data to obtain electric vehicle travel time prediction data; Step 4, calculating to obtain optimal charging power distribution of each electric automobile and charging or non-charging decision of each electric automobile in each time step by utilizing a two-layer optimization strategy combined with a sparrow search algorithm based on the complete electric automobile charging load data and the electric automobile travel time prediction data; The two-layer optimization strategy combining the sparrow search algorithm comprises a first-layer optimization strategy and a second-layer optimization strategy; the first layer optimization strategy comprises the following steps: (1.1) first decision variable Q t first decision variable Q t is the power allocated to the N t vehicle at time t, as follows: ; where Q t j is the power distribution value of vehicle j at time t, j is 1-N t ,N t is the number of vehicles willing to participate in the optimization at time t, , Is the total number of vehicles at time t, The number of vehicles which do not participate in optimization at time t; (1.2) first objective function The first objective function is the total satisfaction of the user at time t, and the formula is as follows: ; where S j is the satisfaction status of vehicle j, When F j is greater than 80%, the single user S j is set to 1 and F j is And (3) with Is used in the ratio of (a), C j is the battery capacity of vehicle j, Is the charge demand of the vehicle j, Is the target SOC of the vehicle j, Is the initial SOC of vehicle j; (1.3) a first model constraint comprising: (1.3.1) the charge amount constraint is: ; Wherein, the After the basic load is deducted and the electric vehicle load is not participated in dispatching, the residential power distribution network can be used for participating in dispatching the residual distributable charging power of the vehicle at the time t, and the formula is as follows: ; In the formula, Is the upper limit of the residential power distribution network; As a predicted value of the base load of the present charge cycle, The method comprises the steps that the total charging load of the electric vehicles which do not participate in dispatching is the total load collection amount corresponding to the vehicle collection which does not participate in dispatching and is obtained by dividing complete electric vehicle charging load data according to user wishes; Is that The average value of the predicted departure time of the number of vehicles participating in the user is calculated as follows: ; In the formula, The predicted departure time of the vehicle j is obtained from travel time prediction data of the electric automobile; Wherein, the The calculation formula of (2) is as follows: ; In the formula, For EV loads that do not participate in the schedule for time i, Is the start time of the present charge cycle, Is the end time of the present charging cycle; ; wherein A not is an EV number sequence which does not participate in scheduling, Is the charging power of vehicle j; (1.3.2) the charging power constraint is: ; (1.3.3) the charging demand constraint is: ; In the formula, For the target SOC for the present charge cycle, Initial SOC for the present charge cycle; (1.4) solving for Q t as a decision variable using a sparrow search algorithm The optimization problem of the objective function and meeting the constraint of the first model is obtained According to Performing optimal charging power distribution on each electric automobile; The second layer optimization strategy comprises the following steps: (2.1) second decision variable The method is characterized by comprising the following steps: ; Wherein w t is the decision variable matrix, Indicating that the vehicle j is charged at time i, Indicating that vehicle j is not charged at time i ; (2.2) Second objective function The second objective function is To participate in the current charging cycle The average value of the user satisfaction of the number of vehicles is as follows: ; ; ; In the formula, Representing a vehicle The amount of charge/charge that can be obtained within a charging time window, which refers to the slave time To its predicted departure time , Acquiring travel time prediction data of the electric automobile; Representing a vehicle Is set according to the target required electric quantity; for vehicles Is provided, wherein, The charging electric quantity of the unit time step or the electric quantity increment corresponding to the charging power; for bicycle satisfaction, determined by threshold rules when Time taking Otherwise take ; (2.3) A second model constraint comprising: (2.3.1) obtaining an energy constraint: ; (2.3.2) distribution amount of electricity constraint: ; In the formula, Is the basic load value at time i, The total charging load of the electric automobile which does not participate in scheduling in the time i; (2.4) solving with a sparrow search Algorithm For decision variables to Is an objective function and meets the optimization problem of the constraint of the second model to obtain According to And obtaining the charging or non-charging decision of each electric automobile in each time step.
2. The method for constructing an electric vehicle charging policy according to claim 1, wherein in step 2, the objective function of the electric vehicle load data restoration model based on structured tensor parallel factorization is as follows: ; In the formula, Is a repaired data tensor for representing predicted charge load data, g r represents an adaptive calculation factor, Is the 1 st factor of the factor matrix, Is the 2 nd factor of the factor matrix, Is the nth factor of the factor matrix; tensors representing original electric vehicle load data; representing the Frobenius norm for calculating the error of the matrix or tensor; An observation portion representing data; Is used for controlling parameters in the tensor decomposition process; is a factor matrix in tensor decomposition used as a low-dimensional representation of the charge load data; Is an error term.
3. The method for constructing the electric vehicle charging strategy according to claim 2, wherein the objective function of the electric vehicle load data restoration model based on the structured tensor parallel factorization is solved by adopting a hierarchical least square method to obtain complete electric vehicle charging load data.
4. The method for constructing the electric vehicle charging strategy according to claim 3, wherein in step 3, the xLSTM-based electric vehicle travel time prediction model comprises an input feature construction module, a mLSTM subnetwork, a sLSTM subnetwork, a feature fusion module and a regression output module; The input feature construction module is used for constructing the data subjected to dimension reduction into an input sequence with the length of T according to the time sequence; the mLSTM subnetwork is used for carrying out time sequence modeling on the input sequence and outputting a first hidden characteristic sequence; The sLSTM subnetwork is used for carrying out time sequence modeling on the input sequence and outputting a second hidden characteristic sequence; the feature fusion module is used for splicing or weighting and fusing the first hidden feature sequence and the second hidden feature sequence to obtain fusion features; And the regression output module is used for outputting the travel time prediction data of the electric automobile based on the fusion characteristics.
5. An electric vehicle charging policy construction system for taking into account user charging behavior, for implementing the electric vehicle charging policy construction method for taking into account user charging behavior as defined in any one of claims 1 to 4, characterized by comprising a data acquisition unit, a data restoration unit, a predictive data calculation unit and a two-layer optimization policy calculation unit; The data acquisition unit is used for acquiring original multidimensional data; The original multidimensional data comprise electric vehicle charging load data, electric vehicle arrival time historical observation data, electric vehicle departure time historical observation data, meteorological data, battery charge state data and target SOC; the data restoration unit is used for restoring the electric vehicle charging load data in the original multidimensional data by utilizing an electric vehicle load data restoration model based on structured tensor parallel factorization to obtain complete electric vehicle charging load data; The prediction data calculation unit is used for carrying out dimension reduction processing on the original multidimensional data to obtain dimension reduced data, and based on the dimension reduced data, predicting by using an electric vehicle travel time prediction model based on xLSTM to obtain electric vehicle travel time prediction data; The two-layer optimization strategy calculation unit is used for calculating and obtaining optimal charging power distribution of each electric automobile and charging or non-charging decision of each electric automobile in each time step by utilizing a two-layer optimization strategy combined with a sparrow search algorithm based on the complete electric automobile charging load data and the electric automobile travel time prediction data.
6. The electric vehicle charging policy building system according to claim 5, wherein the objective function of the electric vehicle load data restoration model based on structured tensor parallel factorization is as follows: ; In the formula, Is a repaired data tensor for representing predicted charge load data, g r represents an adaptive calculation factor, Is the 1 st factor of the factor matrix, Is the 2 nd factor of the factor matrix, Is the nth factor of the factor matrix; tensors representing original electric vehicle load data; representing the Frobenius norm for calculating the error of the matrix or tensor; An observation portion representing data; Is used for controlling parameters in the tensor decomposition process; is a factor matrix in tensor decomposition used as a low-dimensional representation of the charge load data; Is an error term.
7. The electric vehicle charging policy construction system considering the charging behavior of the user according to claim 6, wherein the objective function of the electric vehicle load data restoration model based on the structured tensor parallel factorization is solved by adopting a hierarchical least square method, so as to obtain complete electric vehicle charging load data.
8. The electric vehicle charging strategy construction system considering the charging behavior of the user according to claim 7, wherein the xLSTM-based electric vehicle travel time prediction model comprises an input feature construction module, a mLSTM subnetwork, a sLSTM subnetwork, a feature fusion module and a regression output module; The input feature construction module is used for constructing the data subjected to dimension reduction into an input sequence with the length of T according to the time sequence; the mLSTM subnetwork is used for carrying out time sequence modeling on the input sequence and outputting a first hidden characteristic sequence; The sLSTM subnetwork is used for carrying out time sequence modeling on the input sequence and outputting a second hidden characteristic sequence; the feature fusion module is used for splicing or weighting and fusing the first hidden feature sequence and the second hidden feature sequence to obtain fusion features; And the regression output module is used for outputting the travel time prediction data of the electric automobile based on the fusion characteristics.
9. The electric vehicle charging policy construction system according to claim 8, wherein the two-layer optimization policy combined with the sparrow search algorithm comprises a first-layer optimization policy and a second-layer optimization policy; the first layer optimization strategy comprises the following steps: (1.1) first decision variable Q t first decision variable Q t is the power allocated to the N t vehicle at time t, as follows: ; where Q t j is the power distribution value of vehicle j at time t, j is 1-N t ,N t is the number of vehicles willing to participate in the optimization at time t, , Is the total number of vehicles at time t, The number of vehicles which do not participate in optimization at time t; (1.2) first objective function The first objective function is the total satisfaction of the user at time t, and the formula is as follows: ; where S j is the satisfaction status of vehicle j, When F j is greater than 80%, the single user S j is set to 1 and F j is And (3) with Is used in the ratio of (a), C j is the battery capacity of vehicle j, Is the charge demand of the vehicle j, Is the target SOC of the vehicle j, Is the initial SOC of vehicle j; (1.3) a first model constraint comprising: (1.3.1) the charge amount constraint is: ; Wherein, the After the basic load is deducted and the electric vehicle load is not participated in dispatching, the residential power distribution network can be used for participating in dispatching the residual distributable charging power of the vehicle at the time t, and the formula is as follows: ; In the formula, Is the upper limit of the residential power distribution network; As a predicted value of the base load of the present charge cycle, The method comprises the steps that the total charging load of the electric vehicles which do not participate in dispatching is the total load collection amount corresponding to the vehicle collection which does not participate in dispatching and is obtained by dividing complete electric vehicle charging load data according to user wishes; Is that The average value of the predicted departure time of the number of vehicles participating in the user is calculated as follows: ; In the formula, The predicted departure time of the vehicle j is obtained from travel time prediction data of the electric automobile; Wherein, the The calculation formula of (2) is as follows: ; In the formula, For EV loads that do not participate in the schedule for time i, Is the start time of the present charge cycle, Is the end time of the present charging cycle; ; wherein A not is an EV number sequence which does not participate in scheduling, Is the charging power of vehicle j; (1.3.2) the charging power constraint is: ; (1.3.3) the charging demand constraint is: ; In the formula, For the target SOC for the present charge cycle, Initial SOC for the present charge cycle; (1.4) solving for Q t as a decision variable using a sparrow search algorithm The optimization problem of the objective function and meeting the constraint of the first model is obtained According to Performing optimal charging power distribution on each electric automobile; The second layer optimization strategy comprises the following steps: (2.1) second decision variable The method is characterized by comprising the following steps: ; Wherein w t is the decision variable matrix, Indicating that the vehicle j is charged at time i, Indicating that vehicle j is not charged at time i ; (2.2) Second objective function The second objective function is To participate in the current charging cycle The average value of the user satisfaction of the number of vehicles is as follows: ; ; ; In the formula, Representing a vehicle The amount of charge/charge that can be obtained within a charging time window, which refers to the slave time To its predicted departure time , Acquiring travel time prediction data of the electric automobile; Representing a vehicle Is set according to the target required electric quantity; for vehicles Is provided, wherein, The charging electric quantity of the unit time step or the electric quantity increment corresponding to the charging power; for bicycle satisfaction, determined by threshold rules when Time taking Otherwise take ; (2.3) A second model constraint comprising: (2.3.1) obtaining an energy constraint: ; (2.3.2) distribution amount of electricity constraint: ; In the formula, Is the basic load value at time i, The total charging load of the electric automobile which does not participate in scheduling in the time i; (2.4) solving with a sparrow search Algorithm For decision variables to Is an objective function and meets the optimization problem of the constraint of the second model to obtain According to And obtaining the charging or non-charging decision of each electric automobile in each time step.

Description

Electric vehicle charging strategy construction method and system considering user charging behavior Technical Field The application relates to the technical field of electric vehicle charging, in particular to an electric vehicle charging strategy construction method and system considering user charging behaviors. Background In recent years, the global new energy automobile industry rapidly develops, the holding capacity of electric automobiles greatly rises, and the charging power requirements of the corresponding electric automobiles also obviously increase. This growing trend places heavy load pressures on existing power generation and supply infrastructure. The charging behavior of the current electric automobile user has obvious period centralization characteristic, and the charging time of the current electric automobile user is generally coincident with the peak period height of the residential electricity consumption. The superposition phenomenon is very easy to cause overload operation of the power distribution network, so that the power supply reliability of the power grid can be affected, and interference can be caused to normal production and living electricity utilization of residents. Meanwhile, as the number of electric vehicles continues to increase, the demand for the capacity of the power distribution network is synchronously increased. However, under the technical condition that the capacity of the existing power distribution network is constant, the continuous increase of the number of the electric vehicles directly leads to the fact that part of the electric vehicles cannot complete full charge, and the charging experience and the charging efficiency are greatly reduced. It is particularly noted that the existing residential power distribution network does not fully consider the charging load requirements of the electric vehicle in the initial design stage, so that the power distribution capacity of the electric vehicle is relatively low, and the residential area becomes a high-rise area with incomplete charging. Therefore, how to scientifically and reasonably manage the charging behavior of the electric automobile, and coordinate the contradiction between the charging requirement and the power grid bearing capacity have become the technical problem to be solved currently. Disclosure of Invention Aiming at the technical problems pointed out in the background art, the invention aims to provide an electric vehicle charging strategy construction method and system considering user charging behaviors. In order to achieve the purpose of the invention, the technical scheme provided by the invention is as follows: First aspect The application provides an electric vehicle charging strategy construction method considering user charging behavior, which comprises the following steps: Step1, collecting original multidimensional data; The original multidimensional data comprise electric vehicle charging load data, electric vehicle arrival time historical observation data, electric vehicle departure time historical observation data, meteorological data, battery charge state data and target SOC; Step 2, repairing the electric vehicle charging load data in the original multidimensional data by using an electric vehicle load data repairing model based on structured tensor parallel factor decomposition to obtain complete electric vehicle charging load data; step 3, performing dimension reduction processing on the original multidimensional data to obtain dimension reduced data, and predicting by using an electric vehicle travel time prediction model based on xLSTM based on the dimension reduced data to obtain electric vehicle travel time prediction data; and 4, calculating to obtain optimal charging power distribution of each electric automobile and charging or non-charging decision of each electric automobile in each time step by utilizing a two-layer optimization strategy combined with a sparrow search algorithm based on the complete electric automobile charging load data and the electric automobile travel time prediction data. Second aspect The application provides an electric vehicle charging strategy construction system considering user charging behaviors, which is used for realizing the electric vehicle charging strategy construction method considering the user charging behaviors and comprises a data acquisition unit, a data restoration unit, a predicted data calculation unit and a two-layer optimization strategy calculation unit; The data acquisition unit is used for acquiring original multidimensional data; The original multidimensional data comprise electric vehicle charging load data, electric vehicle arrival time historical observation data, electric vehicle departure time historical observation data, meteorological data, battery charge state data and target SOC; the data restoration unit is used for restoring the electric vehicle charging load data in the original multidimensional data by util