CN115307650A - Electric vehicle charging path planning method based on deep learning - Google Patents
Electric vehicle charging path planning method based on deep learning Download PDFInfo
- Publication number
- CN115307650A CN115307650A CN202210849629.8A CN202210849629A CN115307650A CN 115307650 A CN115307650 A CN 115307650A CN 202210849629 A CN202210849629 A CN 202210849629A CN 115307650 A CN115307650 A CN 115307650A
- Authority
- CN
- China
- Prior art keywords
- charging
- model
- electric automobile
- electric vehicle
- path planning
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3407—Route searching; Route guidance specially adapted for specific applications
- G01C21/3415—Dynamic re-routing, e.g. recalculating the route when the user deviates from calculated route or after detecting real-time traffic data or accidents
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/34—Route searching; Route guidance
- G01C21/3453—Special cost functions, i.e. other than distance or default speed limit of road segments
- G01C21/3469—Fuel consumption; Energy use; Emission aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Automation & Control Theory (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Genetics & Genomics (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
Abstract
The invention provides an electric vehicle charging path planning method based on deep learning, and belongs to the field of electric vehicle path planning. From the perspective of charging station operation, predicting the number of the charging station use piles in a specific area by using a Deep Belief Network (DBN) model, and aiming at providing constraint conditions for path planning of the electric vehicle; starting from the benefit of a user, predicting the average speed of the electric automobile on a specific road by using the LSTM model, taking the prediction result of the LSTM model as the input of the driving energy consumption model of the electric automobile, then establishing a comprehensive energy consumption model of the electric automobile, and further establishing a total travel time model of the electric automobile and a total cost model of the user; and designing an objective function which has the shortest total travel time, the lowest cost of the total cost of the user and the comprehensively optimal total travel time and the total cost of the user, and finally solving the path planning problem under different objectives through an improved genetic algorithm.
Description
Technical Field
The invention relates to the field of electric vehicle path planning, in particular to an electric vehicle charging path planning method based on deep learning.
Background
In the field of electric vehicle path planning research, few scholars combine deep learning prediction models to perform systematic research from electric vehicles, charging stations and users. In addition, aiming at the path planning problem of the electric vehicle, the Dijkstra algorithm is one of the most classical algorithms for solving the path planning, the algorithm traverses the shortest path from the starting point to each point by an expansion method and records the shortest path, and the path finally reaching the end point is recorded as the shortest path. However, the conventional path planning algorithm and the single charging path planning target cannot meet the actual life requirements. Therefore, it is necessary to study the electric vehicle charging path planning.
Disclosure of Invention
Aiming at the defects, the invention provides an electric vehicle charging path planning method based on deep learning. The method mainly comprises the steps of establishing a DBN model after analyzing factors influencing the number of charging piles of the charging station, and predicting the number of the piles used by the charging station to serve as one of path planning constraint conditions. From the perspective of electric automobiles, the method mainly comprises the steps of establishing a traffic network model, analyzing factors influencing the driving speed of the electric automobiles, establishing an LSTM model to predict the speed of the electric automobiles, and further establishing an electric automobile driving energy consumption model. From the perspective of user benefits, a path planning objective function is designed, and the path planning problem is solved by utilizing an improved genetic algorithm.
In order to achieve the purpose, the invention provides the following technical scheme:
an electric vehicle charging path planning method based on deep learning comprises the following steps:
step 2, based on the acquired information, predicting the average speed of the electric automobile at the next moment through the constructed long-short term memory LSTM model, taking the predicted average speed at the next moment as the input of the driving energy consumption model of the electric automobile, and calculating by improving a genetic algorithm to obtain the driving energy consumption; predicting the using quantity of the charging piles at the next moment by constructing a Deep Belief Network (DBN) model, further obtaining the utilization rate of the charging piles, and providing a constraint condition of the utilization rate of the charging piles for improving a genetic algorithm;
and 4, when the electric vehicle user has a charging demand, respectively setting the shortest total travel time, the lowest total cost of the user, and the comprehensively optimal total travel time and the total cost of the user as targets, based on the result predicted in the deep belief network DBN model, further setting a constraint condition, arriving at a charging station from a starting point for charging, arriving at a destination by the charging station, sequentially planning a charging path by improving a genetic algorithm, and obtaining different optimal charging paths according to a target function.
The technical scheme of the invention is further improved as follows: the average vehicle speed prediction method of the long-short term memory LSTM model aims to construct an electric vehicle driving energy consumption model, and comprises the following specific steps:
step 11, aiming at a certain area of a city, establishing a traffic network model, and acquiring information of a starting point and a destination of an electric automobile and position information of a charging station;
step 12, acquiring running speed data of the electric automobile, weather temperature, traffic jam coefficients, intersection density and traffic light density data, determining input and output of a long-short term memory (LSTM) model, and predicting the average speed of the electric automobile at the next moment;
step 13, based on the average speed predicted by the long-term and medium-term memory LSTM in the step 12, taking the average speed as the input of an electric vehicle driving energy consumption model, constructing the electric vehicle driving energy consumption model, and outputting the electric vehicle driving energy consumption model as an electric vehicle comprehensive energy consumption model; the comprehensive energy consumption model of the electric automobile is the sum of an electric automobile driving energy consumption model, an electric automobile load equipment energy consumption model and an electric automobile battery decline model; and further obtaining the total travel time and the total cost of the electric automobile user in the improved genetic algorithm objective function according to the comprehensive energy consumption model of the electric automobile.
The technical scheme of the invention is further improved as follows: the method for predicting the number of charging piles of the charging station of the deep belief network DBN model aims to provide a charging pile utilization rate constraint condition for path planning, namely the constraint condition of path planning in a genetic algorithm is improved, and the prediction method comprises the following steps:
step 21, acquiring information of all charging stations in the traffic network area in the step 1, wherein the information of the charging stations comprises names of the charging stations, the total number of charging piles, charging electric charge, service charge, parking charge and charging pile power for different charging stations; aiming at a specific charging station, acquiring weather temperature, weather conditions, day types, traffic congestion coefficients at every five minutes and the number of piles used by the charging station at every five minutes;
step 22, taking the weather temperature, the weather condition, the day type, the traffic condition at every five minutes and the number of piles used by the charging station at every five minutes as DBN model input information;
step 23, predicting the number of charging piles of the charging station at the next moment at the current moment by using a DBN (deep belief network) model;
step 24, calculating the utilization rate of the charging pile and setting constraint conditions;
step 25, judging whether all the charging stations meet the path planning constraint condition, if so, executing the charging path planning in the step 4 by improving the genetic algorithm in the road network containing the charging stations meeting the constraint condition according to different target functions to obtain different optimal charging paths; if not, the charging station is not considered in the route planning.
The technical scheme of the invention is further improved as follows: the electric automobile total travel time model is as follows:
the electric automobile total travel time model comprises the travel time t 1 Charging waiting time t 2 Charging time t 3 (ii) a The distance between the starting point A and the end point B in the road network is called r i,j (n) N represents different paths, two adjacent nodes in the road network are one path, and m represents the number of the paths; supposing that at the time t, the electric automobile user has a charging demand, and the electric automobile reaches the driving time t of the destination 1 The unit of travel time is h, t 1 The formula is as follows:
in the formula (I), the compound is shown in the specification,
average vehicle speed predicted for the LSTM model;
charging waiting time t of electric automobile 2 Comprises the following steps:
in the formula, H is the number of charging stations in the road network area; x h For the decision variable, X is then selected when charging at the charging station h is selected h =1, otherwise X h ＝0；
Setting that when the charging of the electric automobile reaches a threshold value, the electric automobile can go to a destination; therefore, the charging time t of the electric vehicle 3 Comprises the following steps:
in the formula, E t Represents a threshold value; e s Representing the remaining capacity of the battery; p ch Representing charging power of a charging pile;
in summary, the total travel time T of the electric vehicle is:
T＝t 1 +t 2 +t 3
the technical scheme of the invention is further improved as follows: the electric automobile user total cost model is as follows:
the model of the total cost of the electric automobile user comprises the actual running comprehensive energy consumption cost C 1 Load equipment cost C 2 Charge cost C 3 And the degradation cost C of charging the battery capacity for one time 4 ；
C 3 ＝P ch ·C s ·(t 2 +t 3 )
In the formula, C s Connecting the electric automobile to the electricity price at the charging moment of the power grid; when the air conditioner in the car is started, X 0 Is 1, otherwise is 0; t is t k Indicating a time of arrival at the charging station;
in summary, the total cost C of the electric vehicle users is:
C＝C 1 +C 2 +C 3 +C 4
the technical scheme of the invention is further improved as follows: the charging path objective function is:
in the formula, T is total travel time; c is the total cost of the user; t is avg The average value of the total running time for the electric automobile to go to different charging stations; c avg The average value of the total cost of the electric vehicle for going to different charging stations; lambda 1 And λ 2 As a weight coefficient, when the shortest total travel time is taken as an optimization target, lambda 1 ＝1,λ 2 =0; when the total cost of the user is the lowest as an optimization target, lambda 1 ＝0,λ 2 =1; when the comprehensive optimization of the total travel time and the total cost of the user is taken as an optimization target, lambda 1 ＝1,λ 2 ＝1。
Due to the adoption of the scheme, the technical progress achieved by the invention is as follows: on the premise of considering weather, traffic state and real-time operation state of the charging pile, the shortest travel time and the comprehensive optimal cost of the total cost of the user are taken as a path planning target, and compared with the case that the total travel time is taken as the path planning target, the total travel time is reduced; compared with the path planning target with the lowest total user cost, the total user cost is saved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a LSTM model diagram;
FIG. 2 is a DBN model training structure;
FIG. 3 is a flow chart of an improved genetic algorithm.
Detailed Description
The invention is further illustrated below:
the path method comprises the following steps:
step 2, based on the acquired information, predicting the average speed of the electric automobile at the next moment through the constructed long-short term memory LSTM model, taking the predicted average speed at the next moment as the input of the electric automobile running energy consumption model, and calculating by improving a genetic algorithm to obtain the electric automobile running energy consumption; and predicting the using quantity of the charging piles at the next moment by constructing a Deep Belief Network (DBN) model, further obtaining the utilization rate of the charging piles, and providing a charging pile utilization rate constraint condition for improving a genetic algorithm.
Fill electric pile rate of utilization restraint as follows:
and 3, constructing an electric vehicle total travel time model and an electric vehicle user total cost model based on the result of the LSTM model prediction in the step 2, and further constructing a charging path planning objective function. The objective function is as follows:
in the formula, T is total travel time; c is the total cost of the user; t is a unit of avg The average value of the total running time for the electric automobile to go to different charging stations; c avg The average value of the total cost of the electric vehicle for going to different charging stations; lambda 1 And λ 2 As a weight coefficient, when the shortest total travel time is taken as an optimization target, lambda 1 ＝1,λ 2 =0; when the total cost of the user is the lowest as an optimization target, the lambda is 1 ＝0,λ 2 =1; when the comprehensive optimization of the total travel time and the total cost of the user is taken as an optimization target, lambda 1 ＝1,λ 2 ＝1。
And 4, when the user of the electric automobile has a charging demand, respectively setting a constraint condition of path planning based on a result predicted in the DBN model of the deep belief network by taking the shortest total travel time, the lowest total cost of the user and the comprehensively optimal total travel time and total cost of the user as target functions. And (3) charging from the starting point to the charging station, then, the charging station to the destination, planning the charging path in sequence by improving the genetic algorithm, and obtaining different optimal charging paths according to the objective function.
The average vehicle speed prediction method of the long-short term memory LSTM model aims to construct an electric vehicle driving energy consumption model, and comprises the following specific steps:
and 11, establishing a traffic network model aiming at a certain region of a city, and acquiring the information of the starting point and the destination of the electric automobile and the position information of a charging station. And establishing a directed traffic network model G = (F, R, H) by adopting a graph theory method. Wherein, F = {1,2The variation is a road network node set, and f is the total number of road network nodes selected in the region. H = {1,2,. Eta., H } is the charging station node set, and H is the total number of selected charging station nodes in the area. R is a set of link distances between nodes, wherein the distance between the nodes (i, j | i, j ∈ F) of the road network is R i,j And (4) showing. The distances between the initialized segments are as follows.
In the formula, inf is a road section formed by the node i and the node j and is not directly connected; d ij Is the travel distance between node i and node j.
And step 12, acquiring the running speed data of the electric automobile, the weather temperature, the traffic jam coefficient, the intersection density and the traffic light density data, determining the input and the output of the long-short term memory LSTM model, and predicting the average speed of the electric automobile at the next moment.
Three inputs of the LSTM model for predicting the average vehicle speed are respectively input values x of the network at the current moment t Last time LSTM output value h t-1 Last cell state C t-1 (ii) a The two LSTM outputs are the LSTM output value h at the current moment t And cell state C at the current time t . By forgetting door f t And input gate i t Output gate h t The LSTM model was constructed. Input X (t) and output h of LSTM model 2 (t) is as follows:
in the formula (I), the compound is shown in the specification,
the average speed of the electric automobile at the current moment, T is the weather temperature, and B is the traffic congestionPlugging coefficient, p 1 At a branch road density, ρ 2 In order to achieve the traffic light density,
the next-time average vehicle speed predicted for the LSTM model. Where ρ is 1 At a branch road density, ρ 2 The formula for the traffic light density is as follows:
in the formula, N 1 Number of crossroads, N 2 The number of traffic lights and L the length of the road.
Step 13, based on the average speed predicted by the long-term and short-term memory LSTM in the step 12, taking the average speed as the input of the electric vehicle running energy consumption model, constructing the electric vehicle running energy consumption model, and outputting the electric vehicle running energy consumption model as the electric vehicle comprehensive energy consumption model; the comprehensive energy consumption model of the electric automobile is the sum of an electric automobile driving energy consumption model, an electric automobile load equipment energy consumption model and an electric automobile battery recession model; and according to the comprehensive energy consumption model of the electric automobile, the total travel time and the total cost of the electric automobile user in the target function of the improved genetic algorithm are obtained.
The driving energy consumption model of the electric automobile is as follows:
mechanical power P of electric automobile 1 Comprises the following steps:
in the formula, ρ air Is the air density; m is the vehicle body mass; xi is the quality factor of the electric automobile; a is the electric vehicle acceleration; g is the acceleration of gravity (g =9.8066 m/s) 2 ) (ii) a θ is the road grade value; c. C r Is a rolling resistance parameter;
is the electric vehicle average speed predicted by the LSTM model; a is the frontal windward area of the electric automobile; c. C d Is the coefficient of frictional resistance.
It is assumed that the road gradient value is 0, i.e., θ =0 °, during the driving of the electric vehicle, and the influence of the air resistance and the gradient resistance is not counted. On the premise of considering the energy consumption recovery of regenerative braking, the running motion equation F of the electric automobile 1 Comprises the following steps:
F 1 ＝F f +F b (9)
F b ＝F 2 +F 3 (10)
F b ＝αF 2_max +βF 3_max (11)
wherein α is a weighting coefficient of mechanical power; f 2 Is a frictional braking force; f 2_max Is the maximum friction braking force; f 3 Is the anti-drag braking force provided by the motor; f f Is the rolling resistance; f b Is the acceleration resistance; f 3_max Is the maximum anti-drag braking force; beta is a weighting coefficient of the motor braking force (wherein alpha is more than or equal to 0 and less than or equal to 1, beta is more than or equal to 0 and less than or equal to 1).
Kinetic energy generated by friction braking force of the electric automobile is converted into heat energy, and the heat energy disappears in the air, so that the electric automobile is not beneficial to recycling. Therefore, only F is considered 2_max In the case of =0, the driving energy consumption of the electric vehicle is derived from the braking equation of the anti-drag braking force as follows:
instantaneous power P of electric automobile in driving process 2 Comprises the following steps:
P 2 (t)＝F b ·v(t) (12)
if a braking situation occurs during driving, the energy consumption caused by braking is as follows:
in the formula, v 0 And v 1 Respectively representing the instantaneous speed at the start and the instantaneous speed at the end of braking. The law of conservation of energy is as follows:
ΔE＝∫F 1 ·vdt (14)
let E b ＝∫F b ·vdt (15)
E f ＝∫F f ·vdt＝F f ·l (16)
P 3 ＝Mω (17)
In the formula, E b Representing the kinetic energy generated during braking; e f Representing rolling resistance to generate kinetic energy; l represents a braking distance; m represents motor torque; ω represents the motor angular velocity.
Setting eta 1 For mechanical transmission efficiency, η 2 Is the electric motor generating efficiency, eta 3 For the charging efficiency of the battery, the energy E recovered in the braking process of the electric automobile recov Comprises the following steps:
E recov ＝η 1 η 2 η 3 (ΔE-F f l) (18)
energy consumption E of electric automobile running for d kilometers drive Comprises the following steps:
in conclusion, the electric energy E actually consumed in the running process of the electric automobile actual Comprises the following steps:
E actual ＝E drive +E recov (20)
the energy consumption model of the electric automobile load equipment is as follows:
power P required for maintaining current in-vehicle temperature air Comprises the following steps:
Q 2 ＝Q a +Q d +Q f (21)
Q a ＝δ·A·|T set -T srd | (22)
in the formula, Q 2 Is the heat in the vehicleA load; q a Heat transfer loads for automobiles; q d Is a ventilation load; q f To remove Q a And Q d Other air conditioning heat loads than air conditioning; delta is the heat conduction coefficient of the electric automobile; a is the area of the body of the electric automobile; t is set A temperature set for the air conditioner; t is a unit of srd The temperature of the external environment of the automobile; sigma is the air-conditioning heat load transfer coefficient; tau is the mechanical transmission coefficient of the air conditioner compressor.
When the air conditioner of the electric automobile is used for setting the temperature T set Electric quantity E consumed by air conditioner in operation time t air Comprises the following steps:
E air ＝P air ·t (24)
without considering the external load equipment, the sound equipment is on the road section d ij Consumed electric energy
Comprises the following steps:
the battery capacity fade model is as follows:
in the formula, SOC represents a battery state of charge; CRR represents a battery capacity retention rate; SOC (system on chip) mean Represents an average value of the SOC; Δ SOC represents the amount of change in SOC. Further, the equivalent cycle number EFC can be derived as:
cost per 1% of battery capacity C for the sake of reusing the retired battery 1％ Comprises the following steps:
in the formula, Q 1 Representing the total capacity of the battery.
According to the retired battery price estimation model, the retired battery price p can be estimated u Comprises the following steps:
in the formula, T u Represents the service time of the battery, and the unit is second; p is a radical of n The unit price of the new battery is set as 1.5 yuan/wh; r is e Represents the annual percentage of deterioration, and is set to 20%; DR represents the impression rate, and is set to 6%.
Therefore, the battery degradation cost C of the battery of the electric vehicle for charging once battery Can be calculated from equation (28):
in the formula, N ch Is the total charge and discharge times of the battery; (100-CRR) is the capacity of the battery to decay.
The process of constructing the electric automobile total travel time model and the electric automobile user total cost model is as follows:
the electric automobile total travel time model comprises a travel time t 1 Charging waiting time t 2 Charging time t 3 . The distance between the starting point A and the end point B in the road network is called r i,j (n) N represents different paths, two adjacent nodes in the road network are one path, and m represents the number of the paths. Assuming that at the moment t, the electric automobile user has a charging demand, and the electric automobile reaches the driving time t of the destination 1 The unit of travel time is h.
In the formula (I), the compound is shown in the specification,
the average vehicle speed predicted by the LSTM model.
Charging waiting time t of electric automobile 2 Comprises the following steps:
in the formula, H is the number of charging stations in the road network area; x h For decision variables, X is selected when charging at charging station h is selected h =1, otherwise X h ＝0。
And setting that the electric automobile can go to a destination when the charging reaches a threshold value. Therefore, the charging time t of the electric vehicle 3 Comprises the following steps:
in the formula, E t Represents a threshold value; e s Representing the remaining capacity of the battery; p is ch And indicating the charging power of the charging pile.
In summary, the total travel time T of the electric vehicle is:
T＝t 1 +t 2 +t 3 (34)
the model for the total cost of the user of the electric automobile comprises actual running comprehensive energy consumption cost C 1 Load equipment cost C 2 Charge cost C 3 And the degradation cost C of charging the battery capacity for one time 4 。
C 3 ＝P ch ·C s ·(t 2 +t 3 ) (37)
In the formula, C s Connecting the electric automobile to the electricity price at the charging moment of the power grid; when the air conditioner in the car is turned on, X 0 Is 1, otherwise is 0; t is t k Indicating the moment of arrival at the charging station.
In summary, the total cost C of the electric vehicle users is:
C＝C 1 +C 2 +C 3 +C 4 (40)
a charging pile number prediction method for a charging station of a Deep Belief Network (DBN) model aims to provide a charging pile utilization rate constraint condition for path planning, namely the constraint condition for path optimization in a genetic algorithm is improved, and the prediction method comprises the following steps.
Step 21, acquiring information of all charging stations in the traffic network area in the step 1, wherein the information of the charging stations comprises names of the charging stations, the total number of charging piles, charging electric charge, service charge, parking charge and charging pile power for different charging stations; and acquiring the weather temperature, the weather condition, the day type, the traffic congestion coefficient at the moment of every five minutes and the charging station use pile number at the moment of every five minutes aiming at the specific charging station.
Step 22, taking the number of piles used by the charging station at the weather temperature, the weather condition, the day type, the traffic condition every five minutes and the time every five minutes as DBN model input information; selecting four characteristic data influencing the using quantity of the charging piles as data of a DBN model input layer for model training, wherein the four characteristic data are shown in the following table:
and (3) planning a path by using an improved genetic algorithm, firstly inputting an EV initial position and residual electric quantity, judging whether a charging station exists in the range of the residual mileage of the EV, judging whether the utilization rate of a charging pile is less than 90% if the charging station exists in the range of the residual mileage, initializing algorithm parameters if the utilization rate of the charging pile is less than 90%, carrying out chromosome coding on traffic network nodes and the EV vehicle, and planning an optimal path according to different objective functions. The specific flow is shown in fig. 3 below.
Step 23, predicting the charging pile number of the charging station at the next moment at the current moment by using a DBN (deep belief network) model;
step 24, calculating the utilization rate of the charging pile and setting constraint conditions;
step 25, judging whether all the charging stations meet the path planning constraint condition, if so, executing the genetic algorithm in the step 4, and obtaining different optimal charging paths in the charging station road network containing the constraint condition according to different objective functions; if not, the charging station is not considered in the route planning.
The above-mentioned embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements of the technical solution of the present invention made by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the apparatus of the present invention.
Claims (6)
1. An electric vehicle charging path planning method based on deep learning is characterized by comprising the following steps:
step 1, acquiring the average speed of the electric vehicle, the weather condition, the surrounding traffic jam coefficient, the residual electric quantity, a starting point, a terminal point, the position of a charging station and the use information of the charging pile;
step 2, based on the acquired information, predicting the average speed of the electric automobile at the next moment through the constructed long-short term memory LSTM model, taking the predicted average speed at the next moment as the input of the electric automobile running energy consumption model, and calculating by improving a genetic algorithm to obtain the electric automobile running energy consumption; predicting the using quantity of the charging piles at the next moment by constructing a Deep Belief Network (DBN) model, further obtaining the using rate of the charging piles, and providing a constraint condition of the using rate of the charging piles for improving a genetic algorithm;
step 3, constructing an electric vehicle total travel time model and an electric vehicle user total cost model based on the result of the LSTM model prediction in the step 2, and further constructing a charging path planning objective function;
step 4, when the user of the electric automobile has a charging demand, respectively setting a constraint condition of path planning based on a result predicted in the deep belief network DBN model by taking the shortest total travel time, the lowest total cost of the user, and the comprehensively optimal total travel time and total cost of the user as objective functions; and (4) charging from the starting point to the charging station, then, enabling the charging station to reach the destination, sequentially planning the charging path by improving the genetic algorithm, and obtaining different optimal charging paths according to the objective function.
2. The deep learning-based electric vehicle charging path planning method according to claim 1, wherein the average vehicle speed prediction method of the long-short term memory LSTM model aims to construct an electric vehicle driving energy consumption model, and comprises the following specific steps:
step 11, aiming at a certain area of a city, establishing a traffic network model, and acquiring information of a starting point and a destination of an electric automobile and position information of a charging station;
step 12, acquiring running speed data of the electric automobile, weather temperature, traffic jam coefficients, intersection density and traffic light density data, determining input and output of a long-short term memory (LSTM) model, and predicting the average speed of the electric automobile at the next moment;
step 13, based on the average speed predicted by the long-term and short-term memory LSTM in the step 12, taking the average speed as the input of the electric vehicle running energy consumption model, constructing the electric vehicle running energy consumption model, and outputting the electric vehicle running energy consumption model as the electric vehicle comprehensive energy consumption model; the comprehensive energy consumption model of the electric automobile is the sum of an electric automobile driving energy consumption model, an electric automobile load equipment energy consumption model and an electric automobile battery recession model; and according to the comprehensive energy consumption model of the electric automobile, the total travel time and the total cost of the electric automobile user in the target function of the improved genetic algorithm are obtained.
3. The deep learning-based electric vehicle charging path planning method according to claim 1, wherein the deep belief network DBN model charging station charging pile number prediction method aims to provide a charging pile utilization rate constraint condition for path planning, namely a constraint condition for improving path optimization in a genetic algorithm, and the prediction method comprises the following steps:
step 21, acquiring information of all charging stations in the traffic network area in the step 1, wherein the information of the charging stations comprises names of the charging stations, the total number of charging piles, charging electric charge, service charge, parking charge and charging pile power for different charging stations; acquiring the weather temperature, the weather condition, the day type, the traffic congestion coefficient at every five minutes and the number of the charging station use piles at every five minutes aiming at a specific charging station;
step 22, using the pile numbers of the charging stations at the weather temperature, the weather condition, the day type, the traffic condition at the time of every five minutes and the time of every five minutes as DBN model input information;
step 23, predicting the number of charging piles of the charging station at the next moment at the current moment by using a Deep Belief Network (DBN) model;
step 24, calculating the utilization rate of the charging pile and setting constraint conditions;
step 25, judging whether all the charging stations meet the path planning constraint condition, if so, executing the step 4, and obtaining different optimal charging paths in the charging station road network containing the constraint condition according to different objective functions by improving a genetic algorithm; if not, the charging station is not considered in the route planning.
4. The deep learning-based electric vehicle charging path planning method according to claim 1, wherein the electric vehicle total travel time model is:
the electric automobile total travel time model comprises the travel time t 1 Charging waiting time t 2 Charging time t 3 (ii) a Starting point A and ending point in road networkThe distance between points B is called r i,j (n) N represents different paths, two adjacent nodes in the road network are one path, and m represents the number of the paths; supposing that at the time t, the electric automobile user has a charging demand, and the electric automobile reaches the driving time t of the destination 1 The unit of travel time is h, t 1 The formula is as follows:
in the formula (I), the compound is shown in the specification,
average vehicle speed predicted for the LSTM model;
charging waiting time t of electric automobile 2 Comprises the following steps:
in the formula, H is the number of charging stations in the road network area; x h For the decision variable, X is then selected when charging at the charging station h is selected h =1, otherwise X h ＝0；
Setting that when the charging of the electric automobile reaches a threshold value, the electric automobile can go to a destination; therefore, the charging time t of the electric vehicle 3 Comprises the following steps:
in the formula, E t Represents a threshold value; e s Representing the residual capacity of the battery; p is ch Indicating charging power of the charging pile;
in summary, the total travel time T of the electric vehicle is:
T＝t 1 +t 2 +t 3 。
5. the deep learning-based electric vehicle charging path planning method according to claim 1, wherein the electric vehicle user total cost model is:
the model for the total cost of the user of the electric automobile comprises actual running comprehensive energy consumption cost C 1 Load equipment cost C 2 Charge cost C 3 And the degradation cost C of charging the battery capacity for one time 4
C 3 ＝P ch ·C s ·(t 2 +t 3 )
In the formula, cs is the electricity price of the electric automobile at the moment of connecting the electric automobile to a power grid for charging; when the air conditioner in the car is started, X0 is 1, otherwise, X0 is 0; t is t k Indicating the time of arrival at the charging station;
in summary, the total cost C of the electric vehicle users is:
C＝C 1 +C 2 +C 3 +C 4 。
6. the deep learning-based electric vehicle charging path planning method according to claim 1, wherein the charging path planning objective function is:
in the formula, T is total travel time; c is the total cost of the user; t is a unit of avg The average value of the total running time for the electric automobile to go to different charging stations; c avg The average value of the total cost of the electric vehicle going to different charging stations is obtained; lambda [ alpha ] 1 And λ 2 As a weight coefficient, when the shortest total travel time is taken as an optimization target, lambda 1 ＝1,λ 2 =0; when the total cost of the user is the lowest as an optimization target, lambda 1 ＝0,λ 2 =1; when the comprehensive optimization of the total travel time and the total cost of the user is taken as an optimization target, lambda 1 ＝1,λ 2 ＝1。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210849629.8A CN115307650A (en) | 2022-07-19 | 2022-07-19 | Electric vehicle charging path planning method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210849629.8A CN115307650A (en) | 2022-07-19 | 2022-07-19 | Electric vehicle charging path planning method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115307650A true CN115307650A (en) | 2022-11-08 |
Family
ID=83857462
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210849629.8A Pending CN115307650A (en) | 2022-07-19 | 2022-07-19 | Electric vehicle charging path planning method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115307650A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777181A (en) * | 2023-08-15 | 2023-09-19 | 国网浙江省电力有限公司宁波供电公司 | Method, system and storage medium for adjusting charging pile of derailing type electric automobile |
-
2022
- 2022-07-19 CN CN202210849629.8A patent/CN115307650A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116777181A (en) * | 2023-08-15 | 2023-09-19 | 国网浙江省电力有限公司宁波供电公司 | Method, system and storage medium for adjusting charging pile of derailing type electric automobile |
| CN116777181B (en) * | 2023-08-15 | 2024-04-09 | 国网浙江省电力有限公司宁波供电公司 | Method, system and storage medium for adjusting charging pile of derailing type electric automobile |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Adaptive energy management strategy for fuel cell/battery hybrid vehicles using Pontryagin's Minimal Principle | |
| Xie et al. | Pontryagin’s minimum principle based model predictive control of energy management for a plug-in hybrid electric bus | |
| CN113085665B (en) | Fuel cell automobile energy management method based on TD3 algorithm | |
| Romaus et al. | Optimal energy management for a hybrid energy storage system for electric vehicles based on stochastic dynamic programming | |
| CN110435429B (en) | Dual-motor electric vehicle endurance mileage estimation method fusing energy consumption prediction | |
| CN110775065A (en) | Hybrid electric vehicle battery life prediction method based on working condition recognition | |
| CN107180274B (en) | Typical scene selection and optimization method for electric vehicle charging facility planning | |
| Li et al. | A novel combinatorial optimization algorithm for energy management strategy of plug-in hybrid electric vehicle | |
| Jia et al. | Health-aware energy management strategy for fuel cell hybrid bus considering air-conditioning control based on TD3 algorithm | |
| JP6992419B2 (en) | Electric vehicle cost prediction method, server and electric vehicle | |
| CN115158094A (en) | Plug-in hybrid electric vehicle energy management method based on long-short-term SOC (System on chip) planning | |
| CN114475366B (en) | Fuel cell automobile energy-saving driving method and system based on convex optimization | |
| CN114021391A (en) | Electric vehicle charging load prediction method based on dynamic energy consumption and user psychology | |
| CN115307650A (en) | Electric vehicle charging path planning method based on deep learning | |
| CN114611993A (en) | Urban and rural electric bus dispatching method based on mobile battery pack | |
| Janulin et al. | Energy minimization in city electric vehicle using optimized multi-speed transmission | |
| CN113390430B (en) | Electric vehicle dynamic path planning and charging method for multi-warp stop point trip | |
| Wang et al. | Energy management strategy for fuel cell electric vehicles based on scalable reinforcement learning in novel environment | |
| Styler et al. | Active management of a heterogeneous energy store for electric vehicles | |
| CN112550050B (en) | Electric vehicle charging method and system | |
| Leska et al. | Comparative Calculation of the Fuel–Optimal Operating Strategy for Diesel Hybrid Railway Vehicles | |
| Schellenberg et al. | A computationally inexpensive battery model for the microscopic simulation of electric vehicles | |
| CN115099702B (en) | Electric bus daytime running charging optimization method based on Lagrange relaxation algorithm | |
| CN116358593A (en) | Electric vehicle path planning method, device and equipment considering nonlinear energy consumption | |
| CN115588242A (en) | Energy storage battery testing system and method based on Internet of things |
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 |