[go: up one dir, main page]

CN117350519B - Charging station planning method and system based on new energy passenger car charging demand prediction - Google Patents

Charging station planning method and system based on new energy passenger car charging demand prediction Download PDF

Info

Publication number
CN117350519B
CN117350519B CN202311645475.1A CN202311645475A CN117350519B CN 117350519 B CN117350519 B CN 117350519B CN 202311645475 A CN202311645475 A CN 202311645475A CN 117350519 B CN117350519 B CN 117350519B
Authority
CN
China
Prior art keywords
charging
vehicle
demand
battery
energy
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
Application number
CN202311645475.1A
Other languages
Chinese (zh)
Other versions
CN117350519A (en
Inventor
胡杰
王浩杰
王志红
陈琳
朱雪玲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN202311645475.1A priority Critical patent/CN117350519B/en
Publication of CN117350519A publication Critical patent/CN117350519A/en
Application granted granted Critical
Publication of CN117350519B publication Critical patent/CN117350519B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06315—Needs-based resource requirements planning or analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0204—Market segmentation
    • G06Q30/0205—Market segmentation based on location or geographical consideration
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • Accounting & Taxation (AREA)
  • Tourism & Hospitality (AREA)
  • Game Theory and Decision Science (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Educational Administration (AREA)
  • Primary Health Care (AREA)
  • Artificial Intelligence (AREA)
  • Operations Research (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Quality & Reliability (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

The invention provides a charging station planning method and a charging station planning system based on new energy passenger car charging demand prediction, wherein the method comprises the steps of firstly, collecting and processing operation data of a new energy passenger car, constructing a space-time track diagram, and extracting and obtaining charging and driving related information of the vehicle through the space-time track diagram; then, aiming at the battery capacity attenuation condition of the new energy passenger car, carrying out data reconstruction on the operation data of the new energy passenger car; then, a vehicle journey energy consumption prediction model and a vehicle charging energy prediction model are built, a vehicle energy consumption prediction result and a vehicle battery charging energy prediction result are obtained, and space-time prediction is carried out on the charging demand of the target area; and finally, constructing a double-stage rapid site selection scheme based on a traffic network and charging requirements, obtaining the optimal site selection of the charging station according to a charging requirement space-time prediction result, and completing planning of the specification and the number of the charging piles of the charging station according to the historical charging habits of users and the charging requirement peak value data.

Description

Charging station planning method and system based on new energy passenger car charging demand prediction
Technical Field
The invention belongs to the technical field of new energy, and relates to a planning method and a planning system of a charging station.
Background
The new energy automobile industry has a great breakthrough in the key technical field, the permeability of the new energy automobile is steadily improved, and the matched charging station infrastructure is also required to be gradually perfected along with the gradual improvement of the market share of the new energy automobile.
Although charging stations of new energy automobiles are built on a large scale, the existing charging station infrastructure still has a plurality of problems, such as small number of charging piles, long station searching time, unreasonable distribution of the charging piles, low utilization rate of the charging facilities and the like. It is essential, again because of unreasonable planning of the charging facilities. Therefore, the generation of such problems can be greatly solved by combining computer technology to improve the planning layout of the charging facility.
The existing charging demand prediction mainly starts from charging historical data in a station, and the prediction research of the regional charging demand distribution is performed through historical operation data, so that the method is not suitable for site selection of charging stations without station establishment. Or in combination with existing vehicle inventory, charging demand prediction is achieved from a macroscopic perspective, but this is not applicable to specific charging station planning. The method is also commonly used for researching the regional vehicle charging demand distribution based on the angle of a user travel rule, but the modeling is mostly based on simulation data, few factors are considered, and the final result often has larger deviation from the actual result. In addition, the charging station address selection problem is mainly based on charging requirements, one or two factors in the characteristics of traffic hub and various areas of cities and the travel preference of users are combined, the complexity of the current address selection solving algorithm is high, and the requirement on calculation performance is high.
Disclosure of Invention
In order to solve the problems in the background art, the invention provides a charging station planning method and a charging station planning system based on new energy passenger car charging demand prediction.
The method of the invention comprises the following steps:
firstly, collecting operation data of a new energy passenger car in a target area, removing abnormal data in the operation data to obtain vehicle operation data, constructing a space-time track diagram of the new energy passenger car according to the vehicle operation data and a charging and discharging logic of the new energy passenger car, and extracting charging start time, charging completion time, daily average charging frequency, charging duration and driving path information of the vehicle through the space-time track diagram;
Step two, aiming at the battery capacity attenuation condition of the new energy passenger car, assuming that the battery degradation accords with the cut-off normal distribution, carrying out data reconstruction on the vehicle operation data in the target area to obtain the reconstructed operation data of the new energy passenger car;
Thirdly, constructing a vehicle journey energy consumption prediction model according to a running dataset formed by time information interval of charging starting time, charging completion time data information, daily charging frequency, charging duration, running path information, battery capacity attenuation conditions and reconstruction running data of a new energy passenger vehicle and combining CatBoost algorithm and XGBoost algorithm to obtain a vehicle journey energy consumption prediction result;
Step four, a vehicle charging energy prediction model based on a Stacking fusion frame is constructed, abnormal charging energy values are marked and deleted in the model by using a Bayesian Gaussian mixture algorithm, and a prediction result of the vehicle battery charging energy is obtained;
Step five, combining a vehicle energy consumption prediction result based on travel energy consumption prediction and a vehicle battery charging energy prediction result to perform space-time prediction on the charging demand of the target area;
Step six, constructing a double-stage rapid site selection scheme based on a traffic network and charging demands aiming at a target area according to a charging demand space-time prediction result, wherein the first stage combines thermal distribution to screen out an area which most accords with regional distribution and travel characteristics as a charging station candidate area; the input of the second stage is the charging station candidate region obtained in the first stage, and the K-means algorithm is utilized to carry out iterative calculation on the regional charging demand to obtain the distance average value from the charging station when the charging demand occurs, so as to obtain the optimal address selection of the charging station;
Step seven, firstly setting charge duration and parking duration indexes, extracting historical charge habits of the vehicle according to a space-time trajectory graph, determining a charge type, and planning the specification of a charging pile in a station to which the vehicle belongs; and integrating charging demand peak value data of the target area in an hour unit according to a charging demand space-time prediction result of the target area, and constructing an optimal matching scheme of the number of each power charging pile according to a time node to finish planning of the specifications and the number of the charging piles of the charging station.
In the first step, the operation data of the new energy passenger car include a vehicle number, a data acquisition time, a vehicle state, a charging state, a vehicle speed, a driving mileage, a total voltage, a total current, a vehicle electric quantity, a highest voltage of a battery cell, a lowest voltage of the battery cell, a highest temperature value, a lowest temperature value, a longitude of vehicle running, and a latitude of vehicle running; the abnormal data comprise missing data, jump data and state marking error data; the historical charging habit of the vehicle is the charging type of the vehicle in the space-time track diagram; the charging starting time of the vehicle is the stopping starting time of the vehicle at a charging station in the space-time track diagram; the charging completion time of the vehicle is the stop time of the vehicle at a charging station in the space-time track diagram; the average daily charging frequency of the vehicle is the stay frequency of the vehicle at a charging station in a space-time track diagram; the charging duration time of the vehicle is the total stay time of the vehicle at the charging station in the space-time track diagram; the travel path information of the vehicle is a summary of the longitude of the vehicle travel and the latitude of the vehicle travel in the space-time trajectory graph.
In the second step, the data reconstruction process comprises the steps of firstly completing the identification of the charging attribute to determine whether the charging pile is a public charging pile, then analyzing the battery capacity and the battery capacity attenuation condition of the new energy passenger car, completing the assessment of the battery state of the vehicle by means of a machine learning algorithm and an optimizing algorithm, and completing the identification of the charging attribute and the data reconstruction of the battery state of the vehicle; and the charging attribute identification is to determine whether the attribute of the charging pile is a public charging pile or a private charging pile according to the charging behavior, the position area and the charging power of the passenger car. The data reconstruction of the vehicle battery state is combined with the truncated normal distribution and the battery state which are in accordance with the battery degradation, so that the coupling of the vehicle journey energy consumption prediction model to the battery capacity of the new energy passenger vehicle is completed, and the sample data is more in accordance with the actual vehicle battery state distribution.
The truncated normal distribution is represented as follows:
，
，
wherein, Probability density function as normal distribution,Probability distribution function as normal distribution,AndRespectively represent the mean and standard deviation,AndRespectively represent the upper and lower bounds of the random variable.
Further, in the third step, the vehicle journey energy consumption prediction model includes a path driving state prediction model and a core energy consumption prediction model; the construction method of the path running state prediction model comprises the following steps: constructing various energy consumption factors by utilizing a data mining method, classifying and screening vehicle characteristics by combining a pearson correlation coefficient and a random forest, and dividing the vehicle characteristics into vehicle state characteristics and road condition dynamic characteristics, wherein the vehicle state characteristics are vehicle use static information including the driving mileage and longitude and latitude of a vehicle at the current moment and vehicle current battery information including the current battery capacity, the current battery SOC and the current battery SOH, and the road condition dynamic characteristics are average current and average speed of a battery of the vehicle driving at the current road section; constructing a path running state prediction model of CatBoost algorithm under a Boosting integrated algorithm framework, and predicting to obtain path vehicle energy consumption by taking vehicle state characteristics and road condition dynamic characteristics as inputs; the method for constructing the core energy consumption prediction model comprises the following steps: and calculating actual energy consumption in the running process of the vehicle by using an interval mapping energy consumption estimation method, integrating the output of a path running state prediction model and vehicle state characteristics as input, using XGBoost algorithm under a Boosting integration algorithm frame, establishing a core energy consumption prediction model, and outputting a vehicle energy consumption prediction result.
Further, in the fourth step, the method for constructing the vehicle charging energy prediction model includes:
Firstly, marking and deleting abnormal charge energy values in vehicle charge energy by using a Bayesian Gaussian mixture algorithm, and carrying out distribution statistics on the vehicle charge energy after deleting the abnormal charge energy values, so that the overall distribution of the vehicle charge energy approximately obeys normal distribution, and taking the vehicle charge energy as the input of a vehicle charge energy prediction model;
The calculation of the vehicle charge energy is represented as follows:
，
，
wherein, AndSOC values at the current time and the charging start time,/>, respectivelyFor coulombic efficiency,AndThe battery charging current and the battery capacity are respectively; /(I)Accumulating charging energy for the first k charging moments,AndThe charging voltage and the charging current of the vehicle at the current moment are respectively;
Then selecting a random forest suitable for a linear model in an Embedded embedding method and carrying out feature screening in an L1 regularization mode, wherein the features comprise a battery SOC, a battery capacity value, a battery average temperature value, a battery consistency, a charging power level, a charging average current and a charging starting time, and the screened features comprise the battery SOC, the capacity value and the charging average current;
And using RandomFrost based on Bagging and LGBM based on Boosting as a basic model of Stacking, using Lasso as to be used as a meta model of Stacking, establishing a charging energy prediction model under a Stacking fusion frame, and predicting the battery charging energy of the vehicle.
Further, in the fifth step, the specific method for performing space-time prediction on the charging requirement of the target area is as follows: firstly, a charging behavior decision is completed through a vehicle journey energy consumption prediction model, a charging demand is acquired, and the flow is as follows: firstly, judging according to the current residual SOC value, judging that charging is needed when the SOC value is smaller than a certain threshold value, and judging that potential charging needs exist when the SOC value is larger than the certain threshold value and the parking time is longer than 40 min; meanwhile, the current residual SOC value is input into a stroke energy consumption prediction model, the next stroke residual SOC value is obtained, whether the next stroke starting SOC value is smaller than 0 or not is judged, when the next stroke starting SOC value is smaller than 0, the current residual SOC value is judged to be required to be charged, and otherwise, the current residual SOC value is judged to be not required to be charged;
Determining the charging demand of the vehicle through a charging demand prediction result, and carrying out space-time statistics on the points with the demand to obtain the charging peak period demand time distribution; and respectively carrying out charge demand measurement and calculation on the four charge power levels by combining a charge demand decision result and a charge energy prediction model to obtain a daily peak load charge quantity predicted value.
Further, in the sixth step, the specific method in the first stage is: taking the traveling parking situation and the regional distribution of the vehicles as guidance, evaluating and sorting the regions needing to be built from the aspects of traffic network congestion degree, vehicle parking hot spot distribution and surrounding POI hot spot distribution, and selecting the region which most accords with the regional distribution and traveling characteristics as a charging station candidate region; the specific method of the second stage is as follows: selecting a K-means algorithm in a machine learning clustering algorithm, taking a charging station candidate area obtained in the first stage as initial data of the K-means algorithm, iteratively calculating a demand-to-station distance average value by combining the charging demand prediction result until each station meets 95% of charging demand, ending calculation, and outputting a final site selection result.
The invention provides a charging station planning system based on new energy passenger car charging demand prediction, which comprises a data acquisition module, a data reconstruction module, a vehicle journey energy consumption prediction module, a vehicle charging energy prediction module, a space-time prediction module of charging demand, a charging station address selection module and a charging station planning module.
The data acquisition module is used for acquiring the operation data of the new energy passenger car in the target area and removing abnormal data in the new energy passenger car to obtain vehicle operation data, a space-time track diagram of the new energy passenger car is constructed according to the vehicle operation data and the charge-discharge logic of the new energy passenger car, and the charge start time, the charge completion time, the daily charge frequency, the charge duration and the travel path information of the vehicle are extracted through the space-time track diagram.
The data reconstruction module is used for carrying out data reconstruction on the collected operation data of the new energy passenger car in the target area according to the battery capacity attenuation condition of the new energy passenger car, assuming that the battery degradation accords with the cut-off normal distribution, and obtaining the reconstruction operation data of the new energy passenger car.
The vehicle journey energy consumption prediction module constructs a vehicle journey energy consumption prediction model according to a travel data set formed by time information interval of the charging start time, the charging completion time data information, the average charging frequency, the charging duration time, the travel path information, the battery capacity attenuation condition and the reconstruction operation data of the new energy passenger vehicle, and combines CatBoost algorithm and XGBoost algorithm to obtain a vehicle journey energy consumption prediction result;
The vehicle charging energy prediction module is used for constructing a vehicle charging energy prediction model based on a Stacking fusion frame, and marking and deleting abnormal charging energy values in the model by using a Bayesian Gaussian mixture algorithm to obtain a prediction result of the vehicle battery charging energy.
The space-time prediction module of the charging demand performs space-time prediction on the charging demand of the target area by combining a vehicle energy consumption prediction result based on travel energy consumption prediction and a vehicle battery charging energy prediction result.
The method comprises the steps that an addressing module of the charging station builds a double-stage rapid addressing scheme based on a traffic network and charging requirements aiming at a target area, and the first stage combines thermal distribution to screen out an area which most accords with regional distribution and travel characteristics as a charging station candidate area; and the input of the second stage is the charging station candidate region obtained in the first stage, and the K-means algorithm is utilized to carry out iterative calculation on the regional charging demand to obtain the distance average value from the charging station when the charging demand occurs, so as to obtain the optimal address selection of the charging station.
The charging station planning module firstly sets charging duration and parking duration indexes, extracts historical charging habits of vehicles according to a space-time trajectory graph, determines charging types and plans the specifications of charging piles in a site; and integrating charging demand peak value data of the target area in an hour unit according to a charging demand space-time prediction result of the target area, and constructing an optimal matching scheme of the number of each power charging pile according to a time node to finish planning of the specifications and the number of the charging piles of the charging station.
The invention also proposes a computer device for charging station planning based on new energy passenger vehicle charging demand prediction, comprising a memory, a processor and program instructions stored in the memory for execution by the processor, wherein the processor executes the program instructions to implement the steps in the method and the system described above.
The invention also proposes a computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method described above and the system described above.
Compared with the prior art, the method has the advantages that firstly, the operation data of the new energy passenger car are collected and processed, a space-time track diagram is constructed, and the charging and driving related information of the car is obtained through extraction of the space-time track diagram; then, aiming at the battery capacity attenuation condition of the new energy passenger car, carrying out data reconstruction on the operation data of the new energy passenger car; then, a vehicle journey energy consumption prediction model and a vehicle charging energy prediction model are built, a vehicle energy consumption prediction result and a vehicle battery charging energy prediction result are obtained, and space-time prediction is carried out on the charging demand of the target area; and finally, constructing a double-stage rapid site selection scheme based on a traffic network and charging requirements, obtaining the optimal site selection of the charging station according to a charging requirement space-time prediction result, and completing planning of the specification and the number of the charging piles of the charging station according to the historical charging habits of users and the charging requirement peak value data. According to the method, the influence of the vehicle battery capacity attenuation on the charging requirement is considered, and the data reconstruction is carried out on the running data of the vehicle, so that the prediction accuracy of the charging requirement is greatly improved; according to the invention, the battery state and the actual running state of the vehicle are considered, the vehicle journey energy consumption prediction model and the vehicle charging energy prediction model are respectively constructed, the charging demand of the target area is predicted in a space-time manner based on the prediction result, the accurate prediction of the optimal site selection of the charging station is completed, the specification and the number of the charging piles of the charging station are planned according to the user demand, the station searching distance of the user is shortened to the greatest extent, and the utilization rate of the charging station is improved.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
FIG. 2 is a flow chart for constructing a vehicle trip energy consumption prediction model.
Fig. 3 is a flow chart of a charging behavior decision.
FIG. 4 is a flow chart of a dual stage fast addressing scheme
Detailed Description
In order to make the technical problems, technical schemes and beneficial effects to be solved more clear, the application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
A flow chart of a charging station planning method based on new energy passenger vehicle charging demand prediction is shown in fig. 1, and the specific content is as follows.
Firstly, collecting operation data of a new energy passenger car in a target area, removing abnormal data in the operation data to obtain vehicle operation data, constructing a space-time track diagram of the new energy passenger car according to the vehicle operation data and a charging and discharging logic of the new energy passenger car, and extracting charging start time, charging completion time, average daily charging frequency, charging duration and driving path information of the vehicle through the space-time track diagram.
Specifically, the operation data of the new energy passenger car include a vehicle number, a data acquisition time, a vehicle state, a charging state, a vehicle speed, a driving mileage, a total voltage, a total current, a vehicle electric quantity, a highest voltage of a battery cell, a lowest voltage of the battery cell, a highest temperature value, a lowest temperature value, a longitude of vehicle driving and a latitude of vehicle driving.
The exception data includes missing data, transition data, and status flag error data. The missing data mainly occurs in longitude of vehicle running and latitude data of vehicle running, the jump data mainly occurs in driving mileage data, and the state marking error data mainly occurs in vehicle state and charging state data.
The historical charging habits of the vehicle are the type of charge of the vehicle in the spatiotemporal trace map.
The charging start time of the vehicle is the stop start time of the vehicle at the charging station in the space-time trajectory diagram.
The charging completion time of the vehicle is the stop time of the vehicle at the charging station in the space-time trajectory diagram.
The average daily charging frequency of the vehicle is the number of stops of the vehicle at the charging station in the space-time trajectory graph.
The charging duration of the vehicle is the total length of stay of the vehicle at the charging station in the space-time trajectory diagram.
The travel path information of the vehicle is a summary of the longitude of the travel of the vehicle and the latitude of the travel of the vehicle in the spatiotemporal track diagram.
And secondly, aiming at the battery capacity attenuation condition of the new energy passenger car, assuming that the battery degradation accords with the cut-off normal distribution, carrying out data reconstruction on the vehicle operation data in the target area to obtain the reconstructed operation data of the new energy passenger car.
The data reconstruction process comprises the steps of firstly completing charging attribute identification to determine whether a charging pile is a public charging pile, then analyzing the battery capacity and the battery capacity attenuation condition of a new energy passenger car, completing vehicle battery state assessment by means of a machine learning algorithm and an optimizing algorithm, and completing charging attribute identification and vehicle battery state data reconstruction.
The charging attribute identification is to determine whether the attribute of the charging pile is a public charging pile or a private charging pile according to the charging behavior, the position area and the charging power of the passenger car.
And the data reconstruction of the vehicle battery state is combined with the truncated normal distribution and the battery state which are in accordance with the battery degradation to finish the coupling of the vehicle journey energy consumption prediction model to the battery capacity of the new energy passenger vehicle, so that the sample data is more in accordance with the actual vehicle battery state distribution.
The truncated normal distribution is expressed as follows:
，
，
wherein, Probability density function as normal distribution,Probability distribution function as normal distribution,AndRespectively represent the mean and standard deviation,AndRespectively represent the upper and lower bounds of the random variable.
And thirdly, constructing a vehicle journey energy consumption prediction model according to a travel dataset formed by time information interval according to the charging start time, the charging completion time data information, the average charging frequency, the charging duration, the travel path information, the battery capacity attenuation condition and the reconstruction operation data of the new energy passenger vehicle, and combining CatBoost algorithm and XGBoost algorithm to obtain a vehicle journey energy consumption prediction result.
Specifically, the vehicle journey energy consumption prediction model is composed of a path running state prediction model and a core energy consumption prediction model, and the construction flow chart thereof is shown in fig. 2.
The construction method of the path running state prediction model comprises the following steps: and constructing various energy consumption factors by using a data mining method, and classifying vehicle characteristics by combining a Pearson correlation coefficient (Pearson) and a random forest (RF-OOB) together to divide the vehicle characteristics into vehicle state characteristics and road condition dynamic characteristics. The vehicle state features include the running mileage at the current moment of the vehicle, the vehicle use static information of longitude and latitude and the vehicle current battery information including the current battery capacity, the current battery SOC and the current battery SOH, and the road condition dynamic features refer to the average current and the average speed of the battery of the vehicle running on the current road section. And constructing a path running state prediction model of CatBoost algorithm under the Boosting integrated algorithm framework, and predicting to obtain the path vehicle energy consumption by taking vehicle state characteristics and road condition dynamic characteristics as inputs.
The construction method of the core energy consumption prediction model comprises the following steps: and calculating actual energy consumption in the running process of the vehicle by using an interval mapping energy consumption estimation method, integrating the output of a path running state prediction model and vehicle state characteristics as input, using XGBoost algorithm under a Boosting integration algorithm frame, establishing a core energy consumption prediction model, and outputting a vehicle energy consumption prediction result.
And fourthly, constructing a vehicle charging energy prediction model based on a Stacking fusion frame, marking and deleting abnormal charging energy values in the model by using a Bayesian Gaussian mixture algorithm, and obtaining a prediction result of the vehicle battery charging energy.
Specifically, the construction method of the vehicle charging energy prediction model comprises the following steps:
Firstly, marking and deleting abnormal charge energy values in vehicle charge energy by using a Bayesian Gaussian mixture algorithm, and carrying out distribution statistics on the vehicle charge energy after deleting the abnormal charge energy values, so that the overall distribution of the vehicle charge energy approximately obeys normal distribution, and the overall distribution of the vehicle charge energy approximately obeys the normal distribution and is used as input of a vehicle charge energy prediction model.
The calculation of the vehicle charge energy is expressed as follows:
，
，
wherein, AndSOC values at the current time and the charging start time,/>, respectivelyFor coulombic efficiency,AndThe battery charging current and the battery capacity are respectively; /(I)Accumulating charging energy for the first k charging moments,AndThe charging voltage and the charging current of the vehicle at the present time are respectively.
Then selecting a random forest suitable for a linear model in an Embedded embedding method and carrying out feature screening in an L1 regularization mode, wherein the features comprise a battery SOC, a battery capacity value, a battery average temperature value, a battery consistency, a charging power level, a charging average current and a charging starting time, and the screened features comprise the battery SOC, the capacity value and the charging average current;
And finally, taking RandomFrost based on Bagging and LGBM based on Boosting as a basic model of Stacking, taking Lasso as to be taken as a meta model of Stacking, establishing a charging energy prediction model under a Stacking fusion frame, and predicting the charging energy of the battery of the vehicle.
And fifthly, carrying out space-time prediction on the charging requirement of the target area by combining a vehicle energy consumption prediction result based on travel energy consumption prediction and a vehicle battery charging energy prediction result.
Specifically, the specific method for performing space-time prediction on the charging demand of the target area comprises the following steps: firstly, a charging behavior decision is completed through a vehicle journey energy consumption prediction model, a charging requirement is acquired, and a flow chart of the charging behavior decision is shown in fig. 3, specifically: firstly, judging according to the current residual SOC value, judging that charging is needed when the SOC value is smaller than a certain threshold value, and judging that potential charging needs exist when the SOC value is larger than the certain threshold value and the parking time is longer than 40 min; and simultaneously, inputting the current residual SOC value into a stroke energy consumption prediction model, obtaining the residual SOC value of the next stroke, judging whether the initial SOC value of the next stroke is smaller than 0, and judging that the initial SOC value of the next stroke is required to be charged when the initial SOC value of the next stroke is smaller than 0, otherwise, judging that the initial SOC value of the next stroke is not required to be charged.
Determining the charging demand of the vehicle through a charging demand prediction result, and carrying out space-time statistics on the points with the demand to obtain the charging peak period demand time distribution; and respectively carrying out charge demand measurement and calculation on the four charge power levels by combining a charge demand decision result and a charge energy prediction model to obtain a daily peak load charge quantity predicted value.
Step six, constructing a double-stage rapid site selection scheme based on a traffic network and charging demands aiming at a target area according to a charging demand space-time prediction result, wherein the first stage combines thermal distribution to screen out an area which most accords with regional distribution and travel characteristics as a charging station candidate area; and the input of the second stage is the charging station candidate region obtained in the first stage, and the K-means algorithm is utilized to carry out iterative calculation on the regional charging demand to obtain the distance average value from the charging station when the charging demand occurs, so as to obtain the optimal address selection of the charging station.
A flow chart of the dual stage fast addressing scheme is shown in particular in fig. 4.
The specific method of the first stage is as follows: taking the vehicle travel parking situation and the regional distribution as guidance, evaluating and sorting the regions needing to be built from the aspects of traffic network congestion degree, vehicle parking hot spot distribution and surrounding POI hot spot distribution, and selecting the region which is most in line with the regional distribution and travel characteristics as a charging station candidate region.
The specific method of the second stage is as follows: selecting a K-means algorithm in a machine learning clustering algorithm, taking a charging station candidate area obtained in the first stage as initial data of the K-means algorithm, iteratively calculating a demand-to-station distance average value by combining the charging demand prediction result until each station meets 95% of charging demand, ending calculation, and outputting a final site selection result.
Step seven, firstly setting charge duration and parking duration indexes, extracting historical charge habits of the vehicle according to a space-time trajectory graph, determining a charge type, and planning the specification of a charging pile in a station to which the vehicle belongs; and integrating charging demand peak value data of the target area in an hour unit according to a charging demand space-time prediction result of the target area, and constructing an optimal matching scheme of the number of each power charging pile according to a time node to finish planning of the specifications and the number of the charging piles of the charging station.
The invention also provides a charging station planning system based on the new energy passenger car charging demand prediction, which consists of a data acquisition module, a data reconstruction module, a vehicle journey energy consumption prediction module, a vehicle charging energy prediction module, a charging demand space-time prediction module, a charging station address selection module and a charging station planning module.
The data acquisition module is used for acquiring the operation data of the new energy passenger car in the target area and removing abnormal data in the new energy passenger car to obtain vehicle operation data, a space-time track diagram of the new energy passenger car is constructed according to the vehicle operation data and the charge-discharge logic of the new energy passenger car, and the charge start time, the charge completion time, the daily charge frequency, the charge duration and the travel path information of the vehicle are extracted through the space-time track diagram.
The data reconstruction module is used for carrying out data reconstruction on the collected operation data of the new energy passenger car in the target area according to the battery capacity attenuation condition of the new energy passenger car, assuming that the battery degradation accords with the cut-off normal distribution, and obtaining the reconstruction operation data of the new energy passenger car.
The vehicle journey energy consumption prediction module constructs a vehicle journey energy consumption prediction model according to a travel dataset formed by time information interval of the charging start time, the charging completion time data information, the daily charging frequency, the charging duration, the travel path information, the battery capacity attenuation condition and the reconstruction operation data of the new energy passenger vehicle, and combines CatBoost algorithm and XGBoost algorithm to obtain a vehicle journey energy consumption prediction result.
The vehicle charging energy prediction module is used for constructing a vehicle charging energy prediction model based on a Stacking fusion frame, and marking and deleting abnormal charging energy values in the model by using a Bayesian Gaussian mixture algorithm to obtain a prediction result of the vehicle battery charging energy.
The space-time prediction module of the charging demand performs space-time prediction on the charging demand of the target area by combining a prediction result of the vehicle energy consumption based on the travel energy consumption prediction and a prediction result of the vehicle battery charging energy.
The method comprises the steps that an addressing module of the charging station builds a double-stage rapid addressing scheme based on a traffic network and charging requirements aiming at a target area, and the first stage combines thermal distribution to screen out an area which most accords with regional distribution and travel characteristics as a charging station candidate area; and the input of the second stage is the charging station candidate region obtained in the first stage, and the K-means algorithm is utilized to carry out iterative calculation on the regional charging demand to obtain the distance average value from the charging station when the charging demand occurs, so as to obtain the optimal address selection of the charging station.
The charging station planning module firstly sets charging duration and parking duration indexes, extracts historical charging habits of the vehicle according to the space-time trajectory graph, determines charging types and plans the specifications of charging piles in the station; and integrating charging demand peak value data of the target area in an hour unit according to a charging demand space-time prediction result of the target area, and constructing an optimal matching scheme of the number of each power charging pile according to a time node to finish planning of the specifications and the number of the charging piles of the charging station.
The specific operation method of each module in the system is described in the steps in the method, and is not described herein.
The invention also proposes a computer device for charging station planning based on new energy passenger vehicle charging demand prediction, comprising a memory, a processor and program instructions stored in the memory for execution by the processor, the processor executing the program instructions to implement the steps in the method and the system described above.
The invention also proposes a computer-readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method described above and the system described above.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. The scheme in the embodiment of the application can be realized by adopting various computer languages, such as object-oriented programming language Java, an transliteration script language JavaScript and the like.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the application.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present application without departing from the spirit or scope of the application. Thus, it is intended that the present application also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims (9)

1.基于新能源乘用车充电需求预测的充电站规划方法，其特征在于，包括以下步骤：1. A charging station planning method based on the forecast of charging demand for new energy passenger vehicles, characterized by the following steps: 步骤一、采集目标区域的新能源乘用车的运行数据并剔除其中的异常数据，得到车辆运行数据，依据车辆运行数据结合新能源乘用车的充放电逻辑构建新能源乘用车的时空轨迹图，通过时空轨迹图提取得到车辆的充电类型、充电开始时间、充电完成时间、日均充电频次、充电持续时间、行驶路径信息；Step 1: Collect the operation data of new energy passenger vehicles in the target area and remove abnormal data to obtain vehicle operation data. Based on the vehicle operation data and the charging and discharging logic of new energy passenger vehicles, construct the spatiotemporal trajectory map of new energy passenger vehicles. Extract the vehicle's charging type, charging start time, charging completion time, average daily charging frequency, charging duration, and driving path information from the spatiotemporal trajectory map. 步骤二、针对新能源乘用车的电池容量衰减情况，假设电池退化符合截断式正态分布，对目标区域内的车辆运行数据进行数据重构，得到新能源乘用车的重构运行数据；Step 2: Regarding the battery capacity degradation of new energy passenger vehicles, assuming that the battery degradation follows a truncated normal distribution, the vehicle operation data in the target area is reconstructed to obtain the reconstructed operation data of new energy passenger vehicles. 数据重构过程为:首先完成充电属性辨识确定充电桩是否为公共充电桩，然后分析新能源乘用车的电池容量和电池容量衰减情况，借助机器学习算法和寻优算法，完成对车辆电池状态评估，完成充电属性辨识与车辆电池状态的数据重构；The data reconstruction process is as follows: First, the charging attribute identification is completed to determine whether the charging pile is a public charging pile. Then, the battery capacity and battery capacity decay of the new energy passenger vehicle are analyzed. With the help of machine learning algorithms and optimization algorithms, the vehicle battery status is evaluated, and the data reconstruction of charging attribute identification and vehicle battery status is completed. 所述充电属性辨识是根据乘用车充电行为、位置区域和充电功率，确定充电桩的属性是公共充电桩还是私人充电桩；The charging attribute identification determines whether a charging pile is a public or private charging pile based on the charging behavior, location area, and charging power of the passenger vehicle. 所述车辆电池状态的数据重构结合电池退化符合的截断式正态分布和电池状态，完成车辆行程能耗预测模型对新能源乘用车电池容量的耦合，使样本数据更加贴合真实车辆电池状态分布；The data reconstruction of the vehicle battery status combines the truncated normal distribution of battery degradation with the battery status to complete the coupling of the vehicle trip energy consumption prediction model with the battery capacity of new energy passenger vehicles, making the sample data more consistent with the actual vehicle battery status distribution. 所述截断式正态分布表示如下：The truncated normal distribution is represented as follows: 其中，为正态分布的概率密度函数，为正态分布的概率分布函数，和分别表示均值和标准差，a和b分别表示随机变量的取值上下界；in, Let be the probability density function of a normal distribution. Let be the probability distribution function of a normal distribution. and Let a and b represent the mean and standard deviation, respectively, and let a and b represent the upper and lower bounds of the random variable, respectively. 步骤三、依据车辆的充电开始时间、充电完成时间数据信息、日均充电频次、充电持续时间、行驶路径信息、电池容量衰减情况以及新能源乘用车的重构运行数据组成的时间信息区间化后的行驶数据集，结合CatBoost算法和XGBoost算法构建车辆行程能耗预测模型，得到车辆行程能耗预测结果；Step 3: Based on the time information interval-based driving dataset composed of vehicle charging start time, charging completion time data, average daily charging frequency, charging duration, driving route information, battery capacity degradation, and reconstructed operation data of new energy passenger vehicles, a vehicle trip energy consumption prediction model is constructed by combining the CatBoost algorithm and the XGBoost algorithm to obtain the vehicle trip energy consumption prediction results. 步骤四、构建基于Stacking融合框架的车辆充电能量预测模型，在模型里利用贝叶斯高斯混合算法对异常充电能量值标记并删除，得到车辆电池充电能量的预测结果；Step 4: Construct a vehicle charging energy prediction model based on the Stacking fusion framework. In the model, use the Bayesian-Gaussian mixture algorithm to mark and delete abnormal charging energy values to obtain the prediction results of vehicle battery charging energy. 步骤五、结合基于行程能耗预测的车辆能耗预测结果和车辆电池充电能量的预测结果，对目标区域的充电需求进行时空预测；Step 5: Combine the vehicle energy consumption prediction results based on trip energy consumption prediction and the vehicle battery charging energy prediction results to make a spatiotemporal prediction of the charging demand in the target area. 步骤六、根据充电需求时空预测结果，针对目标区域构建基于交通网络和充电需求的双阶段快速选址方案，第一阶段结合热力分布筛选出最符合地区分布和出行特性的区域作为充电站候选区域；第二阶段的输入为第一阶段得到的充电站候选区域，利用K-means算法对地区充电需求进行迭代计算得到充电需求发生时到充电站的距离均值，得到充电站的最佳选址；Step 6: Based on the spatiotemporal prediction results of charging demand, construct a two-stage rapid site selection scheme based on transportation network and charging demand for the target area. In the first stage, the area that best matches the regional distribution and travel characteristics is selected as the candidate area for charging stations by combining heat distribution. The input of the second stage is the candidate area for charging stations obtained in the first stage. The K-means algorithm is used to iteratively calculate the regional charging demand to obtain the average distance from the charging station when the charging demand occurs, and the optimal site for the charging station is obtained. 步骤七、首先设置充电时长和停车时长指标，根据时空轨迹图提取得到车辆的历史充电习惯，确定充电类型，规划所属站点中充电桩规格；然后根据目标区域的充电需求时空预测结果，以小时为单位整合目标区域充电需求峰值数据，并依照时间节点构建各功率充电桩数量的最优匹配方案，完成对充电站的充电桩规格和数量的规划。Step 7: First, set the charging duration and parking duration indicators. Extract the vehicle's historical charging habits from the spatiotemporal trajectory map, determine the charging type, and plan the specifications of the charging piles in the relevant stations. Then, based on the spatiotemporal prediction results of the charging demand in the target area, integrate the peak charging demand data of the target area on an hourly basis, and construct the optimal matching scheme for the number of charging piles of each power according to the time nodes to complete the planning of the specifications and number of charging piles in the charging station. 2.根据权利要求1所述的基于新能源乘用车充电需求预测的充电站规划方法，其特征在于：所述步骤一中，所述新能源乘用车的运行数据包括车辆编号、数据采集时间、车辆状态、充电状态、车辆车速、行驶里程、总电压、总电流、车辆电量、电池单体最高电压、电池单体最低电压、最高温度值、最低温度值、车辆行驶的经度、车辆行驶的纬度；2. The charging station planning method based on the charging demand forecast of new energy passenger vehicles according to claim 1, characterized in that: in step one, the operating data of the new energy passenger vehicles includes vehicle number, data collection time, vehicle status, charging status, vehicle speed, mileage, total voltage, total current, vehicle power, maximum voltage of battery cells, minimum voltage of battery cells, maximum temperature value, minimum temperature value, longitude of vehicle travel, and latitude of vehicle travel. 所述异常数据包括缺失数据、跳变数据和状态标记错误数据；The abnormal data includes missing data, transition data, and status flag error data; 所述车辆的充电开始时间为时空轨迹图中车辆在充电站的停留开始时间；The charging start time of the vehicle is the start time of the vehicle's stay at the charging station in the spatiotemporal trajectory diagram; 所述车辆的充电完成时间为时空轨迹图中车辆在充电站的停留结束时间；The charging completion time of the vehicle is the end time of the vehicle's stay at the charging station in the spatiotemporal trajectory diagram. 所述车辆的日均充电频次为时空轨迹图中车辆在充电站的停留次数；The average daily charging frequency of the vehicle is the number of times the vehicle stops at the charging station in the spatiotemporal trajectory map; 所述车辆的充电持续时间为时空轨迹图中车辆在充电站的停留总时长；The charging duration of the vehicle is the total time the vehicle stays at the charging station in the spatiotemporal trajectory diagram; 所述车辆的行驶路径信息为时空轨迹图中车辆行驶的经度、车辆行驶的纬度的汇总。The vehicle's travel path information is a summary of the vehicle's longitude and latitude in the spatiotemporal trajectory map. 3.根据权利要求2所述的基于新能源乘用车充电需求预测的充电站规划方法，其特征在于：所述步骤三中，车辆行程能耗预测模型包括构建路径行驶状态预测模型和核心能耗预测模型；3. The charging station planning method based on the charging demand prediction of new energy passenger vehicles according to claim 2, characterized in that: in step three, the vehicle trip energy consumption prediction model includes constructing a path driving state prediction model and a core energy consumption prediction model; 所述路径行驶状态预测模型的构建方法为：利用数据挖掘方法构建多方面能耗因子，结合皮尔逊相关系数和随机森林相结合的方式共同对车辆特征进行分类与筛选，将其分为车辆状态特征和路况动态特征，所述车辆状态特征为包括车辆当前时刻的行驶里程、经纬度的车辆使用静态信息和包括当前电池容量、当前电池SOC、当前电池SOH在内的车辆当下电池信息，所述路况动态特征是指车辆在当前路段行驶的电池平均电流、平均车速；构建Boosting集成算法框架下CatBoost算法的路径行驶状态预测模型，以车辆状态特征和路况动态特征为输入，预测得到路径车辆能耗；The method for constructing the path driving state prediction model is as follows: Multiple energy consumption factors are constructed using data mining methods. Vehicle features are then classified and filtered using a combination of Pearson correlation coefficient and random forest, categorizing them into vehicle state features and road condition dynamic features. The vehicle state features include the vehicle's current mileage, latitude and longitude static information, and current battery information including current battery capacity, current battery SOC, and current battery SOH. The road condition dynamic features refer to the vehicle's average battery current and average vehicle speed while traveling on the current road segment. A path driving state prediction model based on the CatBoost algorithm within the Boosting ensemble algorithm framework is constructed, using the vehicle state features and road condition dynamic features as inputs to predict the vehicle's energy consumption along the path. 所述核心能耗预测模型的构建方法为：利用区间映射能耗估计法计算车辆行驶中的实际能耗、整合路径行驶状态预测模型的输出和车辆状态特征作为输入，使用Boosting集成算法框架下的XGBoost算法，建立核心能耗预测模型，输出车辆能耗预测结果。The core energy consumption prediction model is constructed as follows: the actual energy consumption of the vehicle during driving is calculated using the interval mapping energy consumption estimation method, the output of the integrated path driving state prediction model and the vehicle state characteristics are used as input, and the XGBoost algorithm under the Boosting ensemble algorithm framework is used to establish the core energy consumption prediction model and output the vehicle energy consumption prediction results. 4.根据权利要求1所述的基于新能源乘用车充电需求预测的充电站规划方法，其特征在于：所述步骤四中，车辆充电能量预测模型的构建方法为：4. The charging station planning method based on the prediction of charging demand for new energy passenger vehicles according to claim 1, characterized in that: in step four, the method for constructing the vehicle charging energy prediction model is as follows: 首先利用贝叶斯高斯混合算法对车辆充电能量中异常充电能量值进行标记并删除，并对删除异常充电能量值之后的车辆充电能量进分布统计，使得车辆充电能量整体分布近似服从正态分布，将其作为车辆充电能量预测模型的输入；First, the Bayesian-Gaussian mixture algorithm is used to mark and delete abnormal charging energy values in the vehicle charging energy. Then, the distribution statistics of the vehicle charging energy after deleting abnormal charging energy values are performed to make the overall distribution of vehicle charging energy approximately follow a normal distribution. This is used as the input of the vehicle charging energy prediction model. 所述车辆充电能量的计算表示如下：The calculation of the vehicle charging energy is expressed as follows: 其中，SOC(t)和SOC(t0)分别为当前时刻与开始充电时刻的SOC值，ηc(t)为库伦效率，I(t)和Cn分别为电池充电电流和电池的容量；C_energy为前k个充电时刻累计充电能量，T表示前k个充电时刻的总充电时间，U(t)和I(t)分别是当前时刻的车辆的充电电压和充电电流；Where SOC (t) and SOC (t0) are the SOC values at the current time and the start of charging, respectively; ηc (t) is the coulombic efficiency; I (t) and Cn are the battery charging current and battery capacity, respectively; C_energy is the cumulative charging energy of the first k charging times; T represents the total charging time of the first k charging times; and U (t) and I (t) are the vehicle charging voltage and charging current at the current time, respectively. 然后选用Embedded嵌入法中适合于线性模型的随机森林与L1正则化的方式进行特征筛选，所述特征包括电池SOC、电池容量值、电池平均温度值、电池一致性、充电功率等级、充电平均电流、充电开始时间，从中筛选出的特征包括电池SOC、容量值、充电平均电流；Then, the embedded method, which is suitable for linear models, uses random forest and L1 regularization to select features. The features include battery SOC, battery capacity, battery average temperature, battery consistency, charging power level, average charging current, and charging start time. The features selected from these features include battery SOC, capacity, and average charging current. 将基于Bagging的RandomFrost和基于Boosting的LGBM作为Stacking的基模型，Lasso作为Stacking的元模型，建立Stacking融合框架下的充电能量预测模型，对车辆的电池充电能量进行预测。By using RandomFrost based on Bagging and LGBM based on Boosting as the base models for Stacking, and Lasso as the meta-model for Stacking, a charging energy prediction model under the Stacking fusion framework is established to predict the charging energy of vehicle batteries. 5.根据权利要求1所述的基于新能源乘用车充电需求预测的充电站规划方法，其特征在于：所述步骤五中，对目标区域的充电需求进行时空预测的具体方法为：首先通过车辆行程能耗预测模型完成充电行为决策，获取充电需求，流程为：首先根据当前剩余SOC值进行判断，当SOC值小于某一阈值时，判定为需要充电，当SOC值大于某一阈值且停车时长大于40min时，判定为存在潜在的充电需求；同时将当前剩余SOC值输入到行程能耗预测模型中，获得下一个行程剩余SOC值，判断下一行程起始SOC值是否小于0，当小于0时判定其为需要充电，反之判定为不需要充电；5. The charging station planning method based on the prediction of charging demand for new energy passenger vehicles according to claim 1, characterized in that: in step five, the specific method for spatiotemporal prediction of charging demand in the target area is as follows: firstly, the charging behavior decision is completed through the vehicle trip energy consumption prediction model to obtain the charging demand. The process is as follows: firstly, the current remaining SOC value is judged. When the SOC value is less than a certain threshold, it is determined that charging is required. When the SOC value is greater than a certain threshold and the parking time is greater than 40 minutes, it is determined that there is potential charging demand. At the same time, the current remaining SOC value is input into the trip energy consumption prediction model to obtain the remaining SOC value of the next trip. It is judged whether the starting SOC value of the next trip is less than 0. When it is less than 0, it is determined that charging is required; otherwise, it is determined that charging is not required. 通过充电需求预测结果，完成对车辆充电需求的确定，对有需求的点位进行时空统计，获取充电高峰时期需求时间分布；结合充电需求决策结果和充电能量预测模型，对四个充电功率等级分别进行充电需求量测算，求得日均峰值负荷充电量预测值。Based on the charging demand forecast results, the vehicle charging demand is determined, and spatiotemporal statistics are performed on the locations with demand to obtain the time distribution of demand during peak charging periods. Combining the charging demand decision results and the charging energy prediction model, the charging demand is calculated for each of the four charging power levels to obtain the predicted value of the average daily peak load charging volume. 6.根据权利要求5所述的基于新能源乘用车充电需求预测的充电站规划方法，其特征在于：所述步骤六中，第一阶段的具体方法为：以车辆出行停驻情况与地区分布为导向，从交通网络拥挤程度、车辆停驻热点分布、周边POI热点分布多角度出发，对需要建站的地区进行评估排序，挑选出最符合地区分布与出行特性的区域作为充电站候选区域；6. The charging station planning method based on the prediction of charging demand for new energy passenger vehicles according to claim 5, characterized in that: in step six, the specific method of the first stage is: guided by the vehicle travel and parking situation and regional distribution, from multiple perspectives such as the degree of traffic network congestion, the distribution of vehicle parking hotspots, and the distribution of surrounding POI hotspots, the areas that need to build stations are evaluated and ranked, and the areas that best match the regional distribution and travel characteristics are selected as candidate areas for charging stations. 第二阶段的具体方法为：选用机器学习聚类算法中的K-means算法，将第一阶段得到充电站候选区域作为K-means算法的初始数据，结合充电需求预测结果，迭代计算需求到站距离均值，直至每个站点满足95％的充电需求，结束计算，输出最终的选址结果。The specific method for the second stage is as follows: the K-means algorithm in machine learning clustering is selected, and the candidate charging station areas obtained in the first stage are used as the initial data for the K-means algorithm. Combined with the charging demand prediction results, the average distance from the demand to the station is iteratively calculated until each station meets 95% of the charging demand, the calculation ends, and the final site selection result is output. 7.基于新能源乘用车充电需求预测的充电站规划系统，其特征在于：包括数据采集获取模块、数据重构模块、车辆行程能耗预测模块、车辆充电能量预测模块、充电需求的时空预测模块、充电站的选址模块、充电站规划模块；7. A charging station planning system based on the prediction of charging demand for new energy passenger vehicles, characterized in that: it includes a data acquisition module, a data reconstruction module, a vehicle trip energy consumption prediction module, a vehicle charging energy prediction module, a spatiotemporal prediction module for charging demand, a charging station site selection module, and a charging station planning module. 所述数据采集获取模块用于采集目标区域的新能源乘用车的运行数据并剔除其中的异常数据，得到车辆运行数据，依据车辆运行数据结合新能源乘用车的充放电逻辑构建新能源乘用车的时空轨迹图，通过时空轨迹图提取得到车辆的充电类型、充电开始时间、充电完成时间、日均充电频次、充电持续时间、行驶路径信息；The data acquisition module is used to collect the operation data of new energy passenger vehicles in the target area and remove abnormal data to obtain vehicle operation data. Based on the vehicle operation data and the charging and discharging logic of the new energy passenger vehicles, a spatiotemporal trajectory map of the new energy passenger vehicles is constructed. The charging type, charging start time, charging completion time, average daily charging frequency, charging duration, and driving path information of the vehicle are extracted from the spatiotemporal trajectory map. 所述数据重构模块针对新能源乘用车的电池容量衰减情况，假设电池退化符合截断式正态分布，对采集到的目标区域内新能源乘用车的运行数据进行数据重构，得到新能源乘用车的重构运行数据；数据重构过程为:首先完成充电属性辨识确定充电桩是否为公共充电桩，然后分析新能源乘用车的电池容量和电池容量衰减情况，借助机器学习算法和寻优算法，完成对车辆电池状态评估，完成充电属性辨识与车辆电池状态的数据重构；The data reconstruction module addresses the battery capacity degradation of new energy passenger vehicles. Assuming that battery degradation follows a truncated normal distribution, it reconstructs the operational data of new energy passenger vehicles collected within the target area to obtain reconstructed operational data. The data reconstruction process is as follows: first, charging attribute identification is performed to determine whether the charging pile is a public charging pile; then, the battery capacity and battery capacity degradation of the new energy passenger vehicle are analyzed; and with the help of machine learning algorithms and optimization algorithms, the vehicle battery status is evaluated, thus completing the data reconstruction of charging attribute identification and vehicle battery status. 所述充电属性辨识是根据乘用车充电行为、位置区域和充电功率，确定充电桩的属性是公共充电桩还是私人充电桩；The charging attribute identification determines whether a charging pile is a public or private charging pile based on the charging behavior, location area, and charging power of the passenger vehicle. 所述车辆电池状态的数据重构结合电池退化符合的截断式正态分布和电池状态，完成车辆行程能耗预测模型对新能源乘用车电池容量的耦合，使样本数据更加贴合真实车辆电池状态分布；The data reconstruction of the vehicle battery status combines the truncated normal distribution of battery degradation with the battery status to complete the coupling of the vehicle trip energy consumption prediction model with the battery capacity of new energy passenger vehicles, making the sample data more consistent with the actual vehicle battery status distribution. 所述截断式正态分布表示如下：The truncated normal distribution is represented as follows: 其中，为正态分布的概率密度函数，为正态分布的概率分布函数，和分别表示均值和标准差，a和b分别表示随机变量的取值上下界；in, Let be the probability density function of a normal distribution. Let be the probability distribution function of a normal distribution. and Let a and b represent the mean and standard deviation, respectively, and let a and b represent the upper and lower bounds of the random variable, respectively. 所述车辆行程能耗预测模块依据车辆的充电开始时间、充电完成时间数据信息、日均充电频次、充电持续时间、行驶路径信息、电池容量衰减情况以及新能源乘用车的重构运行数据组成的时间信息区间化后的行驶数据集，结合CatBoost算法和XGBoost算法构建车辆行程能耗预测模型，得到车辆行程能耗预测结果；The vehicle trip energy consumption prediction module uses a time-intervalized driving dataset composed of vehicle charging start time, charging completion time, average daily charging frequency, charging duration, driving route information, battery capacity degradation, and reconstructed operation data of new energy passenger vehicles. It then combines the CatBoost and XGBoost algorithms to construct a vehicle trip energy consumption prediction model and obtain the vehicle trip energy consumption prediction results. 所述车辆充电能量预测模块用于构建基于Stacking融合框架的车辆充电能量预测模型，在模型里利用贝叶斯高斯混合算法对异常充电能量值标记并删除，得到车辆电池充电能量的预测结果；The vehicle charging energy prediction module is used to build a vehicle charging energy prediction model based on the Stacking fusion framework. In the model, the Bayesian-Gaussian mixture algorithm is used to mark and delete abnormal charging energy values to obtain the prediction result of vehicle battery charging energy. 所述充电需求的时空预测模块结合基于行程能耗预测的车辆能耗预测结果和车辆电池充电能量的预测结果，对目标区域的充电需求进行时空预测；The spatiotemporal prediction module for charging demand combines the vehicle energy consumption prediction results based on trip energy consumption prediction and the prediction results of vehicle battery charging energy to perform spatiotemporal prediction of charging demand in the target area. 所述充电站的选址模块针对目标区域构建基于交通网络和充电需求的双阶段快速选址方案，第一阶段结合热力分布筛选出最符合地区分布和出行特性的区域作为充电站候选区域；第二阶段的输入为第一阶段得到的充电站候选区域，利用K-means算法对地区充电需求进行迭代计算得到充电需求发生时到充电站的距离均值，得到充电站的最佳选址；The charging station site selection module constructs a two-stage rapid site selection scheme based on the traffic network and charging demand for the target area. In the first stage, the area that best matches the regional distribution and travel characteristics is selected as the candidate area for the charging station by combining heat distribution. The input of the second stage is the candidate area for the charging station obtained in the first stage. The K-means algorithm is used to iteratively calculate the regional charging demand to obtain the average distance from the charging station when the charging demand occurs, and the optimal site for the charging station is obtained. 所述充电站规划模块首先设置充电时长和停车时长指标，根据时空轨迹图提取得到车辆的历史充电习惯，确定充电类型，规划所属站点中充电桩规格；然后根据目标区域的充电需求时空预测结果，以小时为单位整合目标区域充电需求峰值数据，并依照时间节点构建各功率充电桩数量的最优匹配方案，完成对充电站的充电桩规格和数量的规划。The charging station planning module first sets charging duration and parking duration indicators, extracts the vehicle's historical charging habits based on the spatiotemporal trajectory map, determines the charging type, and plans the specifications of charging piles in the respective stations. Then, based on the spatiotemporal prediction results of charging demand in the target area, it integrates the peak charging demand data of the target area on an hourly basis, and constructs the optimal matching scheme for the number of charging piles of each power according to the time nodes, thus completing the planning of the specifications and number of charging piles in the charging station. 8.基于新能源乘用车充电需求预测的充电站规划的计算机设备，包括存储器、处理器和存储在存储器中可供处理器运行的程序指令，其特征在于：所述处理器执行所述程序指令以实现权利要求1至6中任一项所述方法中的步骤和权利要求7所述的系统。8. A computer device for planning charging stations based on the forecast of charging demand for new energy passenger vehicles, comprising a memory, a processor, and program instructions stored in the memory that can be executed by the processor, characterized in that: the processor executes the program instructions to implement the steps of the method according to any one of claims 1 to 6 and the system according to claim 7. 9.一种计算机可读存储介质，存储有计算机程序，其特征在于：所述计算机程序被处理器执行时实现权利要求1至6中任一项所述方法中的步骤和权利要求7所述的系统。9. A computer-readable storage medium storing a computer program, characterized in that: when the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6 and the system according to claim 7.
CN202311645475.1A 2023-12-04 2023-12-04 Charging station planning method and system based on new energy passenger car charging demand prediction Active CN117350519B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311645475.1A CN117350519B (en) 2023-12-04 2023-12-04 Charging station planning method and system based on new energy passenger car charging demand prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311645475.1A CN117350519B (en) 2023-12-04 2023-12-04 Charging station planning method and system based on new energy passenger car charging demand prediction

Publications (2)

Publication Number Publication Date
CN117350519A CN117350519A (en) 2024-01-05
CN117350519B true CN117350519B (en) 2024-05-17

Family

ID=89365296

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311645475.1A Active CN117350519B (en) 2023-12-04 2023-12-04 Charging station planning method and system based on new energy passenger car charging demand prediction

Country Status (1)

Country Link
CN (1) CN117350519B (en)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118246620B (en) * 2024-02-05 2024-12-20 电小新(北京)科技有限公司 Intelligent black box of energy storage power station and application method thereof
CN118246666B (en) * 2024-03-20 2024-11-22 湖北声通科技股份有限公司 Unmanned minibus fleet dispatching system and method
CN118362906B (en) * 2024-06-19 2024-10-29 中国第一汽车股份有限公司 Battery state parameter prediction method, device, electronic device and storage medium
CN118863333B (en) * 2024-06-24 2024-12-06 北京氢远质投新能源汽车有限公司 New energy vehicle fleet management system and method based on stop point analysis
CN118411062B (en) * 2024-07-04 2024-10-29 南京理工大学 Power distribution network payload prediction method and system considering multiple time-space correlations
CN119863018B (en) * 2024-12-25 2025-08-19 国网辽宁省电力有限公司经济技术研究院 Collaborative planning method and management system for new energy
CN119761782A (en) * 2025-03-06 2025-04-04 惠州市国澳通科技有限公司 AI algorithm-based automobile charging efficiency optimization method and system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10065517B1 (en) * 2016-01-22 2018-09-04 State Farm Mutual Automobile Insurance Company Autonomous electric vehicle charging
CN111064181A (en) * 2019-11-20 2020-04-24 国网浙江省电力有限公司杭州供电公司 Configuration method of power supply and charging station based on spatial schedulable characteristics of charging load
WO2021098352A1 (en) * 2019-11-22 2021-05-27 国网福建省电力有限公司 Active power distribution network planning model establishment method taking into consideration site selection and capacity determination of electric vehicle charging stations
CN113191523A (en) * 2021-01-27 2021-07-30 国电南瑞南京控制系统有限公司 Urban electric vehicle rapid charging demand prediction method and device based on data driving mode and behavior decision theory
WO2022135473A1 (en) * 2020-12-22 2022-06-30 国网上海市电力公司 Method for evaluating acceptance capability of electric vehicle in urban distribution network

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9132742B2 (en) * 2012-02-23 2015-09-15 International Business Machines Corporation Electric vehicle (EV) charging infrastructure with charging stations optimumally sited
WO2019023324A1 (en) * 2017-07-26 2019-01-31 Via Transportation, Inc. Systems and methods for managing and routing ridesharing vehicles
US20220410750A1 (en) * 2021-06-09 2022-12-29 MOEV, Inc. System and method for smart charging management of electric vehicle fleets
CN114943362B (en) * 2022-03-22 2025-03-11 上海电力大学 A fast charging load charging guidance method based on adjustable graded charging service fee

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10065517B1 (en) * 2016-01-22 2018-09-04 State Farm Mutual Automobile Insurance Company Autonomous electric vehicle charging
CN111064181A (en) * 2019-11-20 2020-04-24 国网浙江省电力有限公司杭州供电公司 Configuration method of power supply and charging station based on spatial schedulable characteristics of charging load
WO2021098352A1 (en) * 2019-11-22 2021-05-27 国网福建省电力有限公司 Active power distribution network planning model establishment method taking into consideration site selection and capacity determination of electric vehicle charging stations
WO2022135473A1 (en) * 2020-12-22 2022-06-30 国网上海市电力公司 Method for evaluating acceptance capability of electric vehicle in urban distribution network
CN113191523A (en) * 2021-01-27 2021-07-30 国电南瑞南京控制系统有限公司 Urban electric vehicle rapid charging demand prediction method and device based on data driving mode and behavior decision theory

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
到点式服务模式电动汽车电池需求预测;王森;郭萌;聂规划;徐尚英;代四广;;数学的实践与认识(10);全文 *
城市电动汽车充电站选址规划研究;朱柯羽;;科技展望(08);全文 *
数据驱动的新能源公交车能耗预测;胡杰等;机械科学与技术;全文 *
电动汽车动力电池充电能量的预测方法;胡杰;蔡世杰;黄腾飞;王成;杜常清;;机械科学与技术(06);全文 *

Also Published As

Publication number Publication date
CN117350519A (en) 2024-01-05

Similar Documents

Publication Publication Date Title
CN117350519B (en) Charging station planning method and system based on new energy passenger car charging demand prediction
Liu et al. Data-driven intelligent location of public charging stations for electric vehicles
CN112215427B (en) A method and system for reconstructing vehicle trajectory in the absence of bayonet data
CN112381313B (en) Method and device for determining charging pile address
CN118585852B (en) Garbage classification resource optimization management method based on cloud platform
Xing et al. Modelling driving and charging behaviours of electric vehicles using a data-driven approach combined with behavioural economics theory
CN103745111B (en) Pure electric passenger vehicle continual mileage Forecasting Methodology
CN115063184B (en) Electric vehicle charging demand modeling method, system, medium, equipment and terminal
CN110836675B (en) Decision tree-based automatic driving search decision method
CN103745110B (en) Method of estimating operational driving range of all-electric buses
CN114707292B (en) Voltage stability analysis method for distribution network containing electric vehicles
CN115565376B (en) Vehicle journey time prediction method and system integrating graph2vec and double-layer LSTM
Tian et al. Method for predicting the remaining mileage of electric vehicles based on dimension expansion and model fusion
CN115759462A (en) Charging behavior prediction method and device for electric vehicle user and electronic equipment
CN119004184A (en) Charging planning method and system for long-distance travel of new energy automobile
CN115330267A (en) Layout method, device, equipment and medium for charging and swapping facilities based on demand behavior
De Nunzio et al. A time-and energy-optimal routing strategy for electric vehicles with charging constraints
CN120431731B (en) Main line traffic flow control optimization method based on high-performance traffic simulation intelligent agent
CN115691140A (en) A Method for Analysis and Prediction of Temporal and Spatial Distribution of Vehicle Charging Demand
CN118940905B (en) An operation optimization method, device, equipment and medium for a photovoltaic storage coupled hydrogen production system
CN118031968A (en) A fastest path navigation method, navigation system and application for electric vehicles with nonlinear charging rates
CN113902209A (en) Travel route recommendation method, edge server, cloud server, equipment and medium
CN114358416B (en) Public road network partitioning method, system, equipment and medium based on multi-source traffic data
CN119354226B (en) Path generation method for two-wheeled electric vehicle, electronic device and storage medium
CN119780761B (en) Battery health state evaluation method and device based on vehicle running data

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