A three-phase load balancing method for charging new energy vehicles

Abstract

The invention discloses a three-phase load balancing method for charging new energy vehicles, which belongs to the technical field of three-phase unbalanced optimized commutation. The method specifically comprises the following steps: obtaining the power consumption data of users in a monitoring area per unit time within a monitoring period in real time and generating a curve of the user's power consumption changing with time; extracting the power consumption data at any moment from all the changing curves; selecting any moment and extracting the curve of the power consumption data of all users at that moment changing with unit time, clustering all users according to the changing curve to obtain a plurality of category clusters, and marking each category cluster with features in turn; obtaining the feature marks of the category cluster to which the user belongs at any moment and generating a corresponding feature set; re-clustering the target users according to the feature set, calculating the power consumption mean of any category cluster and dividing the users into three-phase phase sequences, and determining the access phase of any user according to the division result. The invention reduces the three-phase load imbalance of the distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of three-phase unbalanced optimized commutation technology, and in particular to a three-phase load balancing method for charging a new energy vehicle. Background Art

[0002] As traditional energy sources are gradually depleted and environmental pollution becomes increasingly serious, new energy vehicles have become one of the hot spots in the development of the automotive industry. With the rapid development of new energy vehicles, electric vehicles (EV) have gradually become an important part of modern transportation. In order to meet the charging needs of electric vehicles, the installation and use of home charging piles are becoming more and more popular. However, while this trend brings convenience, the access of a large number of electric vehicles will have a great impact on the distribution network.

[0003] In the past, the household electricity load was relatively small and stable, and the three-phase load of the power grid was relatively balanced. Therefore, a fixed access phase was determined in advance for each user to ensure the balance of the three-phase load of the power grid. With the widespread access of home charging piles, since the charging process of electric vehicles is usually unidirectional and high-power, if a large number of home charging piles are concentrated on a certain phase of the power grid, or used simultaneously within a specific time period, it may cause the load of that phase to increase significantly, thereby breaking the originally balanced three-phase power load state, causing fluctuations in the power grid voltage, reducing the operating efficiency of the power grid, and increasing line losses and equipment losses. Summary of the invention

[0004] The purpose of the present invention is to provide a three-phase load balancing method for charging a new energy vehicle to solve the following technical problems:

[0005] In the past, the household electricity load was relatively small and stable, and the three-phase load of the power grid was relatively balanced. Therefore, a fixed access phase was determined in advance for each user to ensure the balance of the three-phase load of the power grid. With the widespread access of home charging piles, since the charging process of electric vehicles is usually unidirectional and high-power, if a large number of home charging piles are concentrated on a certain phase of the power grid, or used simultaneously within a specific time period, it may cause the load of that phase to increase significantly, thereby breaking the originally balanced three-phase power load state, causing fluctuations in the power grid voltage, reducing the operating efficiency of the power grid, and increasing line losses and equipment losses.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A three-phase load balancing method for charging a new energy vehicle comprises the following steps:

[0008] S1, set the monitoring area and monitoring period, obtain the power consumption data of users in the monitoring area per unit time in the monitoring period in real time, and generate a curve Q of the change of user power consumption over time based on the power consumption data;

[0009] S2, select any user and obtain all the change curves generated by the user in the monitoring period and mark them as Q1, Q2, ..., Qn, evenly divide the unit time into several sub-time periods, obtain the corresponding time of the initial points of all sub-time periods and mark them as t1, t2, ..., tm, extract the power consumption data of any initial point corresponding to the time from all the change curves in turn and mark them as Dm'1, ​​Dm'1, ​​..., Dm'n, m' is the initial point time corresponding to the sub-time period and m'∈m, obtain the power consumption data of all initial points corresponding to the time, generate the change curve of the power consumption data of any initial point corresponding to the time with time in turn and mark them as q1, q2, ..., qm;

[0010] S3, selecting a time corresponding to any initial point and extracting a curve of the change in the power consumption data of all users in the monitoring area at that time over time, clustering all users in the monitoring area according to the change curve to obtain a number of category clusters, and marking each category cluster in turn; obtaining the feature marks of the category clusters to which all users belong at any time and generating a corresponding feature set;

[0011] S4, calculate the Euclidean distance I between all feature sets, generate the Euclidean distance set U, set the cluster control radius R, take any feature set as the center, calculate the feature set density P within the control radius R, obtain the mean feature set density within the control radius R of all feature sets, and mark the mean feature set density as MinP; for any feature set, if P is greater than MinP, generate a category cluster with the feature set as the core point, and classify all adjacent feature sets within the radius R into the category cluster to obtain several category clusters; calculate the mean power consumption of any category cluster and divide the users into three-phase phase sequence, and determine the access phase of any user according to the division result.

[0012] As a further solution of the present invention: in S3, the process of marking the category clusters is:

[0013] The number of users included in any category cluster is obtained in sequence, and the category clusters are sorted in descending order according to the number of users. Each category cluster is marked according to the arrangement order to obtain the feature mark Bm'x' of any category cluster, where x'∈X, X is the number of all category clusters corresponding to the moment.

[0014] As a further solution of the present invention: In S4, the process of obtaining the Euclidean distance set U is:

[0015] Sorting feature sets , k represents the number of feature sets, m is the number of feature labels in any feature set, and the Euclidean distance I between each two feature sets is calculated respectively. One feature set is marked as P and the other feature set is marked as t. The calculation formula is as follows:

[0016] ;

[0017] Among them, y P ,y t Represent the feature set P and feature set t and the corresponding feature vectors respectively, sort the distance data from each feature set to other feature sets, and get .

[0018] As a further solution of the present invention: in S4, the process of setting the cluster control radius R is:

[0019] The sum of each Euclidean distance data in the set U is obtained to obtain u, and the control radius R is obtained based on the value u. The calculation formula is as follows:

[0020] ;

[0021] ;

[0022] Among them, u is the sum of all Euclidean distance data values ​​in set U.

[0023] As a further solution of the present invention: the process of calculating the feature set density within the control radius R is:

[0024] P=i / (πR 2 );

[0025] Among them, i is the number of feature sets within the control radius R.

[0026] As a further solution of the present invention: in the above S4, if there is a non-core point feature set whose density P within the control radius R is also greater than MinP, the category cluster generated by the non-core point is merged with the original category cluster to generate several category clusters.

[0027] As a further solution of the present invention: in S4, the specific process of dividing the three-phase sequence is:

[0028] Select any category cluster and obtain all users in the category cluster, calculate the average power consumption per unit time of any user during the monitoring period and sum them up to get the total power load W per unit time, set the phase standard value W / 3, use the exhaustive method to combine the phase sequence of users in the category cluster, calculate the deviation value of each phase load in any phase sequence combination and the standard load value W / 3, calculate the evaluation score of each phase sequence combination based on the deviation value, and select the phase sequence combination with the smallest evaluation score as the phase sequence allocation of the current category cluster.

[0029] As a further solution of the present invention: the specific calculation process of the evaluation score is:

[0030] Evaluation score = (Aw-W / 3)² + (Bw-W / 3)² + (Cw-W / 3)²;

[0031] Among them, Aw is the load of phase A allocated to any phase sequence combination, Bw is the load of phase B allocated to the phase sequence combination, and Cw is the load of phase C allocated to the phase sequence combination.

[0032] Beneficial effects of the present invention:

[0033] The present invention obtains the electricity consumption data of users within a monitoring period and generates all change curves generated by users within the monitoring period, and extracts the electricity consumption value at the same time from all change curves to obtain the sub-change curve of the user at each time. Clustering the sub-change curves of all users at the same time can obtain the electricity consumption habits of the users at each time. It can be understood that compared with the traditional full-day load curve, since the sub-load curve focuses on the electricity consumption behavior at a specific time rather than the average electricity consumption of the whole day, it can filter out the data interference caused by daily fluctuations, thereby more accurately reflecting the long-term electricity consumption habits of users, and the length of the sub-load curve is much smaller than the length of the initial load curve, which greatly reduces the actual calculation amount, and obtains all users at any time. The characteristic tags of the category cluster to which they belong are engraved and the corresponding feature set is generated; by calculating the Euclidean distance between the feature sets and setting the clustering control radius according to the Euclidean distance, all users can be divided into several representative category clusters, and by dividing the three-phase phase sequence of the representative category clusters and determining the access phase of any user according to the division result, it can be understood that each user has different electricity usage habits at different times. Therefore, the characteristic labels of the users' different electricity usage habits at different times are obtained and a feature set is generated. Users with the same electricity usage habits can be obtained by clustering the feature sets, and the average electricity consumption of users with the same electricity usage habits is calculated and divided into phase sequences, thereby reducing the imbalance of the three-phase load of the distribution network and improving the stability of the distribution network operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention will be further described below in conjunction with the accompanying drawings.

[0035] Figure 1 It is a flow chart of a three-phase load balancing method for charging a new energy vehicle according to the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] See also Figure 1 As shown, the present invention is a three-phase load balancing method for charging a new energy vehicle, comprising the following steps:

[0038] S1, set the monitoring area and monitoring period, obtain the power consumption data of users in the monitoring area per unit time in the monitoring period in real time, and generate a curve Q of the change of user power consumption over time based on the power consumption data;

[0039] S2, select any user and obtain all the change curves generated by the user in the monitoring period and mark them as Q1, Q2, ..., Qn, evenly divide the unit time into several sub-time periods, obtain the corresponding time of the initial points of all sub-time periods and mark them as t1, t2, ..., tm, extract the power consumption data of any initial point corresponding to the time from all the change curves in turn and mark them as Dm'1, ​​Dm'1, ​​..., Dm'n, m' is the initial point time corresponding to the sub-time period and m'∈m, obtain the power consumption data of all initial points corresponding to the time, generate the change curve of the power consumption data of any initial point corresponding to the time with time in turn and mark them as q1, q2, ..., qm;

[0040] S3, selecting a time corresponding to any initial point and extracting a curve of the change in the power consumption data of all users in the monitoring area at that time over time, clustering all users in the monitoring area according to the change curve to obtain a number of category clusters, and marking each category cluster in turn; obtaining the feature marks of the category clusters to which all users belong at any time and generating a corresponding feature set;

[0041] S4, calculate the Euclidean distance I between all feature sets, generate the Euclidean distance set U, set the cluster control radius R, take any feature set as the center, calculate the feature set density P within the control radius R, obtain the mean feature set density within the control radius R of all feature sets, and mark the mean feature set density as MinP; for any feature set, if P is greater than MinP, generate a category cluster with the feature set as the core point, and classify all adjacent feature sets within the radius R into the category cluster to obtain several category clusters; calculate the mean power consumption of any category cluster and divide the users into three-phase phase sequence, and determine the access phase of any user according to the division result.

[0042] The present invention obtains the electricity consumption data of users within a monitoring period and generates all change curves generated by users within the monitoring period, and extracts the electricity consumption value at the same time from all change curves to obtain the sub-change curve of the user at each time. Clustering the sub-change curves of all users at the same time can obtain the electricity consumption habits of the users at each time. It can be understood that compared with the traditional full-day load curve, since the sub-load curve focuses on the electricity consumption behavior at a specific time rather than the average electricity consumption of the whole day, it can filter out the data interference caused by daily fluctuations, thereby more accurately reflecting the long-term electricity consumption habits of users, and the length of the sub-load curve is much smaller than the length of the initial load curve, which greatly reduces the actual calculation amount, and obtains all users at any time. The characteristic tags of the category cluster to which they belong are engraved and the corresponding feature set is generated; by calculating the Euclidean distance between the feature sets and setting the clustering control radius according to the Euclidean distance, all users can be divided into several representative category clusters, and by dividing the three-phase phase sequence of the representative category clusters and determining the access phase of any user according to the division result, it can be understood that each user has different electricity usage habits at different times. Therefore, the characteristic labels of the users' different electricity usage habits at different times are obtained and a feature set is generated. Users with the same electricity usage habits can be obtained by clustering the feature sets, and the average electricity consumption of users with the same electricity usage habits is calculated and divided into phase sequences, thereby reducing the imbalance of the three-phase load of the distribution network and improving the stability of the distribution network operation.

[0043] It is worth noting that the present invention predicts the electricity consumption behavior of the entire user, so the feature set is the feature mark of the category cluster to which the user belongs at all moments within a unit time, that is, the electricity consumption behavior of each user is similar every day, and then according to the average of the user's electricity consumption in a day, the users with the same electricity consumption behavior can be divided into high-power users and low-power users, so as to dynamically adjust the user access phase. In real time, in order to further reduce the three-phase load imbalance problem of the distribution network, the number of category clusters at all moments included in the feature set can be reduced accordingly, and then the step S4 is repeated, that is, the electricity consumption behavior of each user in a certain time period is similar, and then according to the average of the user's electricity consumption in a certain time period, the users with the same electricity consumption behavior can be divided into high-power users and low-power users, so as to dynamically adjust the user access phase, thereby achieving the optimal phase division for each time period, thereby reducing the three-phase load imbalance of the distribution network and improving the stability of the distribution network operation.

[0044] In another preferred embodiment of the present invention, in S3, the process of marking the category clusters is:

[0045] The number of users included in any category cluster is obtained in sequence, and the category clusters are sorted in descending order according to the number of users. Each category cluster is marked according to the arrangement order to obtain the feature mark Bm'x' of any category cluster, where x'∈X, X is the number of all category clusters corresponding to the moment.

[0046] In another preferred embodiment of the present invention, in S4, the process of obtaining the Euclidean distance set U is:

[0047] Sorting feature sets , k represents the number of feature sets, m is the number of feature labels in any feature set, and the Euclidean distance I between each two feature sets is calculated respectively. One feature set is marked as P and the other feature set is marked as t. The calculation formula is as follows:

[0048] ;

[0049] Among them, y P ,y t Represent the feature set P and feature set t and the corresponding feature vectors respectively, sort the distance data from each feature set to other feature sets, and get .

[0050] In another preferred embodiment of the present invention, in S4, the process of setting the cluster control radius R is:

[0051] The sum of each Euclidean distance data in the set U is obtained to obtain u, and the control radius R is obtained based on the value u. The calculation formula is as follows:

[0052] ;

[0053] ;

[0054] Among them, u is the sum of all Euclidean distance data values ​​in set U.

[0055] In another preferred embodiment of the present invention, the process of calculating the feature set density within the control radius R is:

[0056] P=i / (πR 2 );

[0057] Among them, i is the number of feature sets within the control radius R.

[0058] In another preferred embodiment of the present invention, the S4 further includes merging the category cluster generated by the non-core point with the original category cluster to generate several category clusters if the density P within the control radius R of the non-core point feature set is also greater than MinP.

[0059] In another preferred embodiment of the present invention, in S4, the specific process of dividing the three-phase sequence is as follows:

[0060] Select any category cluster and obtain all users in the category cluster, calculate the average power consumption per unit time of any user during the monitoring period and sum them up to get the total power load W per unit time, set the phase standard value W / 3, use the exhaustive method to combine the phase sequence of users in the category cluster, calculate the deviation value of each phase load in any phase sequence combination and the standard load value W / 3, calculate the evaluation score of each phase sequence combination based on the deviation value, and select the phase sequence combination with the smallest evaluation score as the phase sequence allocation of the current category cluster.

[0061] In another preferred embodiment of the present invention, the specific calculation process of the evaluation score is:

[0062] Evaluation score = (Aw-W / 3)² + (Bw-W / 3)² + (Cw-W / 3)²;

[0063] Among them, Aw is the load of phase A allocated to any phase sequence combination, Bw is the load of phase B allocated to the phase sequence combination, and Cw is the load of phase C allocated to the phase sequence combination.

[0064] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A three-phase load balancing method for charging new energy vehicles, characterized in that: The following steps are involved: S1, set the monitoring area and monitoring period, obtain the power consumption data of users in the monitoring area per unit time in the monitoring period in real time, and generate a curve Q of the change of user power consumption over time based on the power consumption data; S2, select any user and obtain all the change curves generated by the user in the monitoring period and mark them as Q1, Q2, ..., Qn, evenly divide the unit time into several sub-time periods, obtain the corresponding time of the initial points of all sub-time periods and mark them as t1, t2, ..., tm, extract the power consumption data of any initial point corresponding to the time from all the change curves in turn and mark them as Dm'1, ​​Dm'1, ​​..., Dm'n, m' is the initial point time corresponding to the sub-time period and m'∈m, obtain the power consumption data of all initial points corresponding to the time, generate the change curve of the power consumption data of any initial point corresponding to the time with time in turn and mark them as q1, q2, ..., qm; S3, selecting a time corresponding to any initial point and extracting a curve of the change in the power consumption data of all users in the monitoring area at that time over time, clustering all users in the monitoring area according to the change curve to obtain a number of category clusters, and marking each category cluster in turn; obtaining the feature marks of the category clusters to which all users belong at any time and generating a corresponding feature set; S4, calculate the Euclidean distance I between all feature sets, generate the Euclidean distance set U, set the cluster control radius R, take any feature set as the center, calculate the feature set density P within the control radius R, obtain the mean of the feature set density within the control radius R of all feature sets, and mark the mean of the feature set density as MinP; for any feature set, if P is greater than MinP, generate a category cluster with the feature set as the core point, and classify all adjacent feature sets within the radius R into the category cluster to obtain several category clusters; calculate the mean power consumption of any category cluster and divide the users into three-phase phase sequence, and determine the access phase of any user according to the division result; The specific calculation process for setting the cluster control radius R is: The sum of each Euclidean distance data in the set U is obtained to obtain u, and the control radius R is obtained based on the value u. The calculation formula is as follows: Among them, u is the sum of all Euclidean distance data values ​​in set U; The specific calculation process of calculating the feature set density P within the control radius R is: P = i / (πR2); Among them, i is the number of feature sets existing within the control radius R; The specific process of three-phase phase sequence division is: Select any category cluster and obtain all users in the category cluster, calculate the average power consumption per unit time of any user in the monitoring period and sum them up to obtain the total power load W per unit time, set the phase standard value W / 3, use the exhaustive method to combine the phase sequence of users in the category cluster, calculate the deviation value of each phase load in any phase sequence combination and the standard load value W / 3, calculate the evaluation score of each phase sequence combination based on the deviation value, and select the phase sequence combination with the smallest evaluation score as the phase sequence allocation of the current category cluster; Evaluation score = (Aw-W / 3) 2 +(Bw-W / 3) 2 +(Cw-W / 3) 2 ; Wherein, Aw is the load corresponding to phase A of any phase sequence combination, Bw is the load corresponding to phase B of the phase sequence combination, and Cw is the load corresponding to phase C of the phase sequence combination.

2. A three-phase load balancing method for charging a new energy vehicle according to claim 1, characterized in that: In S3, the process of labeling the category clusters is as follows: The number of users included in any category cluster is obtained in sequence, and the category clusters are sorted in descending order according to the number of users. Each category cluster is marked according to the arrangement order to obtain the feature mark Bm'x' of any category cluster, where x'∈X, X is the number of all category clusters corresponding to the moment.

3. A three-phase load balancing method for charging a new energy vehicle according to claim 1, characterized in that: In S4, the process of obtaining the Euclidean distance set U is: Sort the feature set Y l＝1...k ={y1,y2,y3...y m }, k represents the number of feature sets, m is the number of feature labels in any feature set, and the Euclidean distance I between each two feature sets is calculated respectively, one of the feature sets is marked as P, and the other feature set is marked as t. The calculation formula is as follows: Among them, yP, yt represent the feature set P and feature set t and the corresponding feature vectors respectively. The distance data from each feature set to other feature sets are sorted to obtain U l＝1...k ={I1,I2,I3...I k-1 }.

4. A three-phase load balancing method for charging a new energy vehicle according to claim 1, characterized in that: In the above S4, if there is a non-core point feature set whose density P within the control radius R is also greater than MinP, the category cluster generated by the non-core point is merged with the original category cluster to generate several category clusters.

Images