Electric vehicle short-time charging route prediction method based on environmental factors

Abstract

The invention belongs to the technical field of charging loads, and particularly relates to an electric vehicle short-time charging route prediction method based on environmental factors, which comprises the following steps: constructing a traffic road network model; collecting traffic flow data, and performing denoising processing on the traffic flow data through an index moving average method; calculating the correlation of the de-noised alternating current flow data by adopting a Spearman correlation coefficient formula; performing normalization processing on the data traffic flow data according to the correlation, and inputting the data traffic flow data into the prediction model to obtain the traffic flow of the next time period; constructing a flow-density-speed relationship according to the predicted traffic flow, and constructing a road impedance model; obtaining the shortest travel time path of the electric vehicle according to the road impedance model, and calculating the mileage power consumption of the electric vehicle; processing the mileage power consumption by adopting a Monte Carlo method to obtain a prediction result of the space-time distribution of the charging load of the electric vehicle; determining a charging route of the electric vehicle according to the prediction result; according to the invention, charging load prediction is carried out on the electric vehicle with characteristics in the mountain city region, so that the travel convenience of the user is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of electric vehicle charging load, and in particular relates to a method for predicting an electric vehicle short-time charging route based on environmental factors. Background Art

[0002] Under the dual pressure of energy and environmental protection, electric vehicles are developing rapidly around the world, with huge clean and environmental benefits. Electric vehicles have been gradually widely used, and the application of electric vehicles will greatly meet the requirements of today's society for environmental awareness and effectively save the consumption of fossil fuels. Electric vehicles will become an important means of transportation in the future. Due to the intermittent, random and uncertain nature of electric vehicles, the uncertainty of electric vehicle charging behavior has brought a huge impact on the power system. If large-scale electric vehicles are charged disorderly (randomly connected to the power grid for charging), it will affect the stability of the power grid, resulting in peak superposition, which may threaten the operation of the power grid.

[0003] When studying the charging load of electric vehicles, it is easy to ignore the weather factors and traffic conditions on the road. Traffic flow will show randomness due to changes in weather factors, but there are few studies on weather at present. At the same time, many data sets for traffic flow prediction are based on plain cities. Mountain cities refer to cities located and built in mountainous areas. They are constrained by topography and form urban spatial structures that are different from plain cities. Therefore, the impact of terrain on traffic flow should be considered, and some traffic sections with similar spatial structures to mountain cities should be selected as data sets to make the model prediction effect more accurate. The traffic conditions on the road will have a great impact on the energy consumption of electric vehicles when moving, that is, it will affect the spatiotemporal load of electric vehicles. At the same time, the traffic conditions on the road will also have a great impact on the results of the orderly scheduling of electric vehicles. At the same time, due to the vast territory, complex terrain and numerous mountainous cities in my country, the traditional electric vehicle load prediction methods for plain cities ignore the impact of terrain on the driving path and driving power consumption of electric vehicle users, resulting in poor applicability of these methods in cities with special terrains. In addition, travel temperature also affects the energy consumption of electric vehicles during driving. Summary of the invention

[0004] In order to solve the problems existing in the above prior art, the present invention proposes a method for predicting the short-time charging route of electric vehicles based on environmental factors, which includes: collecting traffic flow data, denoising the traffic flow data by exponential moving average method; using Spearman correlation coefficient formula to calculate the correlation of the denoised AC flow data; normalizing the traffic flow data according to the correlation, inputting the normalized data into the BP-LSTM-Attention prediction model to obtain the traffic flow in the next time period; constructing the flow-density-speed relationship according to the predicted traffic flow; and calculating the traffic flow-density-speed relationship according to the traffic flow-density-speed relationship. Construct a road impedance model; obtain the travel time path of the tram according to the road impedance model, and calculate the mileage power consumption of the tram; use the Monte Carlo method to process the mileage power consumption and obtain the prediction result of the spatiotemporal distribution of the tram charging load; construct a tram link model, which consists of multiple charging pile nodes; use the optimized similarity calculation formula to calculate the similarity between the charging pile node and the tram node, and determine the link between the two nodes based on the similarity; use the comprehensive similarity formula to calculate the comprehensive similarity of the link based on the prediction results of the spatiotemporal distribution of the link and the tram charging load; recommend the optimal driving route from the tram node to the charging pile node based on the comprehensive similarity.

[0005] Beneficial effects of the present invention:

[0006] The present invention provides preliminary preparation for subsequent exploration of how to give full play to the dispatching potential of electric vehicles in mountainous cities by predicting the charging load of electric vehicles with characteristics of mountainous urban areas; the present invention constructs a road impedance model according to the flow-density-speed relationship, obtains the shortest travel time path of the tram according to the road impedance model, and calculates the mileage power consumption of the tram, and finally uses the Monte Carlo method to analyze the mileage power consumption of the tram, thereby improving the accuracy of the prediction results of the spatiotemporal distribution of the tram charging load. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 It is a Monte Carlo simulation flow chart of the present invention;

[0008] Figure 2 It is a training flow chart of the BP-LSTM-Attention prediction model of the present invention;

[0009] Figure 3 It is the overall flow chart of the present invention. DETAILED DESCRIPTION

[0010] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0011] A method for predicting short-term charging routes of electric vehicles based on environmental factors, such as Figure 3 As shown, the method includes: constructing a traffic network model; collecting traffic flow data, and denoising the traffic flow data by exponential moving average method; using Spearman correlation coefficient formula to calculate the correlation of the denoised AC flow data; normalizing the data traffic flow data according to the correlation, and inputting the normalized data into the BP-LSTM-Attention prediction model to obtain the traffic flow in the next time period; constructing the flow-density-speed relationship according to the predicted traffic flow; constructing a road impedance model according to the flow-density-speed relationship; obtaining the travel time path of the tram according to the road impedance model, and calculating the mileage power consumption of the tram; using Monte Carlo method to process the mileage power consumption to obtain the prediction result of the spatiotemporal distribution of the tram charging load; constructing a tram link model, which is composed of multiple charging pile nodes; using the optimized similarity calculation formula to calculate the similarity between the charging pile node and the tram node, and determining the link between the two nodes according to the similarity; using the comprehensive similarity formula to calculate the comprehensive similarity of the link according to the prediction results of the spatiotemporal distribution of the link and the tram charging load; and recommending the optimal driving route from the tram node to the charging pile node according to the comprehensive similarity.

[0012] In actual road traffic networks, most of them are in a mesh structure, so graph theory is used to model the traffic network. The expression is:

[0013]

[0014] Where G represents the traffic network of the mountainous city; J represents the set of all nodes in the graph G; E represents the set of all road sections in the graph G; W is the set of road section weights, that is, road resistance; K represents the time period set, dividing the whole day into m time periods; Q is the traffic volume on the road; V is the vehicle speed on the road; θ is the road section slope of the mountainous city road; and n represents the number of nodes.

[0015] E is the set of road segments in the road network, and its adjacency matrix D = (d ij ) n×n Describes the adjacency relationship between nodes and edges in the road network. Its elements are defined as:

[0016]

[0017] In the formula, It represents the road weight between node i and node j, which can be quantified by weights such as road section length and travel time. Inf means that there is no direct road connection between the two nodes.

[0018] The neighbor matrix can be expressed as:

[0019]

[0020] The Spearman Correlation Coefficient is used to study the correlation between two groups of variables. The calculation formula of the Spearman Correlation Coefficient is:

[0021]

[0022]

[0023] Where n is the sample size; (U i ,V i ) are the ranks of the two variables.

[0024] Real-time dynamic traffic flow is a complex system with high uncertainty and randomness. Its changes are often affected by various random factors such as weather, road environment, travel cost and travel time. Although the dynamic change process of traffic flow is complex and constantly fluctuating, there are still certain regularities and periodicities within it, which can be revealed through in-depth mining.

[0025] In the process of collecting traffic flow data, due to the complex and changeable external environment, the collected data may contain outliers. In order to ensure the smooth modeling of traffic flow data, the influence of noise on the original data should be removed. The exponential moving average method (EMA) is selected to denoise the original traffic flow data, and the change trend of traffic flow data is retained to the greatest extent without increasing the complexity of the algorithm.

[0026] EMA t =αX t +(1-α)EMA t-1

[0027] Among them, EMA t is the exponential moving average at time t, α is the smoothing factor, which is a constant between 0 and 1 that determines the weight of the latest observation, X t is the value at time t.

[0028] Traffic flow will change to a certain extent with the change of region and road section. The Spearman correlation coefficient formula is used to calculate the correlation between the current location, current time and traffic flow. The Spearman correlation coefficient is selected to construct the correlation matrix, and the significance test is used to evaluate the strength of the correlation between the two data sets, so as to explore the strong correlation between weather data such as temperature, humidity, visibility, wind direction, rain and snow weather on short-term traffic flow.

[0029] Traffic flow data, geographic information, weather data, etc. are constructed into a geographical and weather factor traffic flow time series matrix according to the observed time series as model data input.

[0030] Since the attributes in the traffic and weather data used have different dimensional standards, which may affect the accuracy of the subsequent prediction model, it is necessary to constrain the values ​​of each attribute to the same value range. Therefore, the traffic flow data is normalized and min-max standardization is used for data normalization.

[0031]

[0032] By adopting the BP-LSTM prediction model based on the attention mechanism, the BP model with back-propagation capability is combined with the LSTM model that is good at mining potential time series features to capture the long-term dependency between data. The Attention mechanism is introduced to reasonably allocate weights according to the strength of the correlation between geographic data, weather data and traffic flow data, and finally a BP-LSTM-Attention prediction model is constructed.

[0033] In this embodiment, the BP-LSTM-Attention prediction model is used to process the input data, including: screening the collected relevant traffic flow data, and constructing a weather factor traffic flow time series matrix according to the time series; normalizing the weather factor traffic flow time series matrix to obtain a standard data matrix; sampling the standard data matrix using a sliding window method to obtain a time series; preprocessing the time series; inputting the preprocessed sequence into the LSTM layer to obtain the dynamic features of the time series; inputting the dynamic features of the time series into the hidden layer for feature enhancement; inputting all enhanced dynamic features into the Attention layer for weight fusion to obtain fused dynamic features; inputting the fused dynamic features into the classification prediction layer to obtain classification prediction results.

[0034] The weather factor traffic flow time series matrix is:

[0035]

[0036] Among them, m represents the number of dimensions, X 1 , X2 , ..., X m They are traffic flow data, atmospheric temperature, relative humidity, visibility, wind direction and other weather data of different dimensions, t 1 , t 2 , ..., t n represents the corresponding observation time. Since the dimension units of data of different dimensions are inconsistent, this will greatly reduce the training efficiency and prediction accuracy of the model. The data is normalized so that the input of the model is a normalized standard data matrix.

[0037] Through the obtained standardized data samples, a training sequence is established using the sliding window method. First, the time series matrix in the above formula is converted into:

[0038]

[0039] Next, the shift() function is used to slide the window, and n=1 is set as the length of the sliding window. According to the time sequence of the data, it is input into the prediction model in batches for prediction. Finally, after repeated learning and training of a large amount of data, the traffic flow of 24 time periods per day is combined with the feature quantity, and after feature screening, the traffic flow prediction model is established.

[0040] The structure of the model includes LSTM layer, hidden layer and Attention layer.

[0041] LSTM layer: First, build the LSTM layer. The input data (preprocessed sequence data) is directly fed into the LSTM layer, which captures the dynamic characteristics of the time series through its internal memory units and forget gates, input gates, output gates and other mechanisms.

[0042] Hidden layer: As needed, additional fully connected layers (Dense layers) or more LSTM layers can be added after the LSTM layer to further extract features or increase the nonlinear capabilities of the model.

[0043] Attention layer: The output of the LSTM layer (usually the hidden state of the last time step or the hidden state of all time steps) is sent to the Attention layer. The function of the Attention layer is to assign different weights according to the importance of the hidden state of each time step, so as to focus on those time steps that have a greater impact on the prediction results.

[0044] like Figure 2 As shown, the training process of the model uses the back propagation algorithm (BP) to train the entire model. During the training process, the weights of the model are updated by the mean absolute error MAE, root mean square error RMSE and determination coefficient (R2).

[0045] The traffic flow of urban roads is predicted and the parameter Q is obtained. In order to evaluate the performance of the prediction model in traffic flow prediction, the mean absolute error MAE, root mean square error RMSE and determination coefficient (R2) are used as evaluation indicators. The mathematical expressions are as follows:

[0046]

[0047]

[0048]

[0049] Among them, y i represents the predicted value, represents the true value of the traffic flow at time i, and m represents the total number of predictions.

[0050] The impact of the special terrain of mountainous cities on vehicle speed is mainly reflected in the road slope and road speed limit, namely:

[0051] V=V max ×e -αθ

[0052] Where α is the adjustment coefficient, which indicates the influence of road slope on speed; V max Designed for the maximum speed limit on mountainous urban roads.

[0053] Speed ​​v, flow Q, and density K are three parameters used to describe traffic flow characteristics. These three parameters are interrelated and together describe the dynamic characteristics of traffic flow. The relationship between the three traffic flow parameters is shown in the formula:

[0054] Q=VK

[0055] Traffic density refers to the number of vehicles per unit length of road within a specific period of time, which can reflect the traffic density and congestion of the road section.

[0056]

[0057] Where K represents the vehicle density; N is the number of vehicles in the road section; and L represents the length of the road section.

[0058] The relationship between velocity and density is:

[0059]

[0060] Where, v represents the speed of the electric vehicle; V f represents the free flow speed; K j represents the blocking density.

[0061] According to the traffic condition classification criteria of the city, the road saturation s is used to represent the traffic condition of the road, where the traffic condition is shown in Table 1.

[0062] Table 1 Traffic conditions

[0063]

[0064] When electric vehicles travel in a road network, the road impedance usually consists of two parts: section impedance and node impedance. Due to the differences in the traffic capacity between road intersections and sections, section impedance and node impedance models corresponding to different saturation s can be established according to their respective traffic capacities.

[0065] The road section impedance model includes: there is a certain relationship between the road section travel time of the vehicle during the driving process and the road's traffic capacity. In addition, weather factors such as changes in climate temperature will affect the vehicle's driving speed, thereby affecting the vehicle's travel time on the road. Considering the climate factors, the road section impedance formula corresponding to different saturations is:

[0066]

[0067]

[0068]

[0069] In the formula, K c is the climate correction coefficient, which can reflect the impact of climate temperature on the vehicle travel time; T c is the real-time climate temperature, T max and T min The upper and lower limits of the suitable temperature range are 20℃ and 25℃ respectively; R vij (t) is the road section impedance, that is, the actual travel time of the road section considering the influence of temperature; t 0 is the travel time under free flow; α and β are impedance influencing factors; in formula (16), s represents the saturation of the road section; Q is the traffic flow of the road section; C D The traffic capacity of the road section.

[0070] Node impedance model. In an urban traffic network, the delay of vehicles at road intersections is called node impedance, which is mainly determined by the capacity limit of the intersection and the control of traffic lights.

[0071]

[0072] In the formula, C vi (t) is the node impedance; c is the signal period; λ is the green light signal time ratio at the node; q is the vehicle arrival rate of the road section.

[0073] Impedance model of mountainous urban roads. Impedance is also called travel time, which is composed of road travel time and intersection delay time in the urban traffic network. After considering the characteristics of mountainous urban roads, the calculation formula of the improved road impedance model is as follows. The improved road impedance model reflects the impact of climate temperature and special mountain terrain on vehicle driving in the traffic network, which is closer to the actual situation.

[0074]

[0075] In the formula, θ is the slope of the mountainous urban road. The influence of the road slope angle on the travel time is mainly achieved by changing the driving speed of the electric vehicle. On the road section with a large slope, the vehicle usually travels at a lower speed, and the vehicle may change speed frequently, which further increases the travel time and thus changes the user's travel time and space parameters.

[0076] In this embodiment, obtaining the shortest travel time path of the tram includes: in the unique terrain of the mountainous city, the traditional Floyd algorithm based on horizontal distance is not enough to capture the authenticity of the EV's travel time in the road network, so an improved Floyd algorithm is used to replace the traditional road distance weight with the actual travel time of the electric vehicle in the mountainous city to determine the shortest travel time between each node. The improved algorithm steps include:

[0077] Step 1: Initialize the adjacency matrix w (0) , set the initial time cost between i and j to If points i and j are not directly connected, then T total =∞;

[0078] Step 2: Construct w (1) , insert the first intermediate node between i and j to optimize the time cost, calculate

[0079] Step 3: Construct w (2) , insert a second node between the two points and calculate

[0080] Step 4: Construct w (n) ,calculate Then (n) represents the shortest time cost between i and j after traversing all nodes, w (n) It forms the shortest time cost matrix between each node.

[0081] In this embodiment, the electric vehicle's unit mileage power consumption uses an energy consumption factor model to calculate the unit mileage power consumption of the EV when it travels on different roads at a certain speed. The expression is:

[0082]

[0083] Where V k is the vehicle's speed, which is affected by the road traffic conditions and the road grade speed limit; k Represents the power consumption per unit mileage of an EV when traveling on a k-level road, where k includes first-level, second-level, third-level, and fourth-level roads.

[0084] At present, there are few studies on the impact of travel temperature on electric vehicle energy consumption in the prediction of electric vehicle charging load. The impact of temperature change on EV energy consumption is mainly on batteries and air conditioners. The correlation between EV energy consumption per unit mileage and ambient temperature is specifically expressed in the following formula.

[0085]

[0086] Where T represents the ambient temperature of the trip, in degrees Celsius (℃); C(T) is the power consumption per unit mileage at different temperatures. Table 2 gives the values ​​of the coefficients in the model.

[0087] Table 2 Values ​​of coefficients in the model

[0088]

[0089] When the ambient temperature is around 20℃, the energy consumption of EV reaches the lowest, so the proportional coefficient m of energy consumption change at different temperatures is established based on 20℃. T , the power consumption ratio at temperature T is as follows:

[0090]

[0091] Taking into account the impact of traffic conditions and ambient temperature on power consumption, the power consumption per unit mileage of EV can be obtained, and its expression is:

[0092]

[0093] In this embodiment, the impact of terrain on power consumption includes: In view of the frequent ups and downs of roads in mountainous cities, when analyzing the actual power consumption of EVs when driving in mountainous urban road networks, the roads should be divided into multiple sections, each of which has its own unique slope and curvature. The actual distance between any two points needs to take into account the ups and downs, and is no longer just a simple straight-line distance. The actual distance between two points is obtained by decomposing the road between the two points into multiple sections and calculating the distance of each section to obtain the actual cumulative distance. The formula is as follows:

[0094]

[0095] In the formula, is the actual spatial distance between points i and j after considering the unique terrain of mountainous cities; d c for The actual spatial distance of segment c in is:

[0096]

[0097] In the formula, x c and h c They represent the horizontal distance and vertical height difference of the cth section of road respectively.

[0098] Therefore, considering the impact of the traffic network impedance of mountainous cities on travel time, temperature, road conditions, etc., the power consumption of electric vehicles on mountainous city roads is as follows:

[0099]

[0100] In the formula, a i,j It is the efficiency coefficient of electric vehicles to overcome gravity or energy recovery in unit relative height, with the unit of m / kWh. i,j It can be shown by the following formula.

[0101]

[0102] When an electric car is driving on an uphill road, it needs to overcome gravity and do work. i,j is the climbing coefficient a c When an electric vehicle is traveling on a downhill section, a portion of the energy can be recovered due to the braking state, so a i,j is the energy recovery coefficient a d .

[0103] In this embodiment, the prediction of the spatiotemporal distribution of charging load includes: based on the analysis of EV user behavior, the probability distribution of the spatiotemporal variables of user travel behavior can be obtained as the input of the Monte Carlo method. When simulating user charging demand, it is necessary to comprehensively consider the influence of external factors and spatiotemporal variables and alternately sample spatiotemporal variables.

[0104] In this embodiment, if Figure 1As shown in the figure, the Monte Carlo method is used to process the mileage power consumption, including: constructing the probability distribution of user travel behavior in each time period in the future through the Monte Carlo method, generating a variety of potential charging scenarios through extensive random simulation operations, summarizing all charging scenarios, and obtaining the probability distribution of load; planning the user's travel path according to the load probability distribution, using the Floyd algorithm to plan the user's travel path with the shortest driving time as the goal, obtaining the travel time of each road based on the speed-flow-density model, and updating its spatiotemporal variables and charge state after the EV arrives at the destination, and calculating the travel power consumption based on the duration and mileage of the trip, combined with the impact of terrain and temperature on energy consumption; updating the remaining power according to the travel power consumption; analyzing the user's charging needs, obtaining the spatiotemporal distribution of each electric vehicle and the information at each destination, when the vehicle arrives at a destination, judging whether the user needs to charge based on the remaining power, if charging is required, determining the charging time based on the residence time and power situation, calculating the charging load of the electric vehicle at each destination, i.e., each functional area, and continuing to drive if charging is not required.

[0105] After travel demand is generated, the user's travel route needs to be planned. The Floyd algorithm is used to plan the user's travel route with the shortest travel time as the goal. The travel time of each road is obtained based on the speed-flow-density model. After the EV arrives at the destination, its spatiotemporal variables and charge state are updated.

[0106] The charging probability formula is:

[0107]

[0108] In the formula, Indicates whether the EV will be charged at destination s, 1 indicates that the user has a charging demand, and 0 indicates that the user will not charge; remaining mileage is the power consumption per unit mileage of electric vehicles in mountainous cities, taking into account the influence of road conditions and temperature; L s,s+1 It is the actual driving distance of the user's next trip in the mountainous urban road network.

[0109] When an EV user arrives at each destination, the travel power consumption will be calculated based on the duration and mileage of the trip, combined with the impact of terrain and temperature on energy consumption, and then the remaining power will be updated to estimate the charging load during the stop at the destination and then integrated into the charging demand of the corresponding area. After each trip to the destination, the initial charge state of the EV user when arriving at the current destination can be obtained based on the mileage and driving path of the electric vehicle and the established mountain city power consumption model. The expression is:

[0110]

[0111] In the formula, Indicates the amount of electricity required to reach the current destination and leave the previous location; SOC (T d_0 ) is the power consumption at the first trip, which is related to the user’s behavior habits; E max The maximum power of EV, which is related to battery parameters; D d and ε d N is the distance traveled by the electric vehicle and the power consumption per kilometer under the corresponding road conditions; S-1,S Represents the total number of road segments from the start point to the end point planned based on the shortest travel time.

[0112] Assume that if an EV user chooses to charge at a certain location, the user immediately accesses the charging facility upon arrival at the location and continues charging until the battery reaches the predetermined ideal power or maximum stay time. The charging time is:

[0113]

[0114] Where η is the charging efficiency of EV; P s represents the EV charging power level at time s; t stay is the dwell time of EV at a destination; E exp It is the user's expected power value, usually 1.

[0115] The charge state of the EV user when leaving the current destination. If the EV chooses not to charge after making a charging decision at the current destination, the charge state when leaving is the same as when arriving at the current destination; if the EV is charged at the destination, the charge state is selected until the electric vehicle is fully charged or leaves the current destination.

[0116]

[0117] In the formula represents the vehicle's residence time; η represents the charging efficiency; P s represents the charging power level at s.

[0118] By analyzing the charging needs of users, we can obtain the spatial and temporal distribution of each electric vehicle and the information at each destination, including arrival time, residence time, departure time and charge level at each time, so as to calculate the charging load of electric vehicles at each destination, that is, each functional area. When the vehicle arrives at a destination, the user will make a charging decision based on the remaining power. If charging is required, the charging time will be determined based on the residence time and power level. The charging load at a certain area f can be expressed as:

[0119]

[0120] In the formula, Indicates the charging status of the electric vehicle at time t, if it is in the charging state, it is 1, if it is not in the charging state, it is 0; N f (t) is the number of electric vehicles in the functional area at time t; is the total charging load of electric vehicles in the functional area at time t.

[0121] Construct an electric vehicle link model, which consists of multiple charging pile nodes; that is, recommend suitable charging stations for electric vehicles, and incorporate load forecasts, time-of-use electricity prices, charging service fees, and passenger source forecasts as node internal attributes into the overall network topology. During the research process, electric taxis and charging stations are abstracted as nodes in the network, and the encounter records between nodes are converted into links in the network, so as to achieve the prediction and recommendation of links that may be formed in the future.

[0122] The tram link model consists of two types of nodes: tram nodes and charging station nodes. The definitions and attributes of the nodes are as follows: tram node (v i ∈V) represents each tram running in the city. Position P i : represents the tram v i The location coordinates at time slot t are usually expressed in longitude and latitude. Current load forecast L j : Reflects the degree of charging demand of the vehicle, mainly used to assess whether the vehicle needs to be given priority for charging.

[0123] Current load forecast L i The expression is:

[0124]

[0125] Among them, C max is the maximum battery capacity of the electric vehicle, C i (t) is the tram v i The remaining battery capacity at time slot t, is the current battery power consumption ratio, reflecting the charging demand of the electric vehicle; T i (t) is the tram v i The current waiting time at time slot t is used to evaluate the possible time cost of the charging station; β 1 It is an adjustment factor used to weigh the impact of remaining power and waiting time.

[0126] Estimated user D i Indicates the potential passenger demand in the area where the vehicle is located. It is used to evaluate the passenger carrying potential of the area within a certain period of time and affects whether the vehicle chooses to charge.

[0127] D i =α 1 ·H i (t)+α 2 ·Si (t)+α 3 ·P i (t)

[0128] Among them, H i (t) is the historical user boarding data, indicating the historical user demand for the tram at the current location. i (t) is the crowd density in the area where the previous vehicle is located, which can be obtained through real-time ride application data or crowd monitoring data; P i (t) is the predicted impact of special factors such as weather and holidays, which is usually predicted by statistical models or machine learning methods; α 1 , α 2 , α 3 It is an adjustment coefficient used to balance the impact of different factors.

[0129] Charging station node: represents each charging station in the city. Each charging station node contains the following attributes:

[0130] Position P j :Indicates charging station c j The fixed position coordinates of .

[0131] Current load forecast L j :Indicates charging station c j The current load forecast reflects the busyness of the charging station and is mainly used to assess the congestion of the charging station.

[0132]

[0133] Where: N j (t) is the charging station c j The current number of vehicles in the queue at time slot t; N max is the maximum service capacity of the charging station, that is, the maximum number of vehicles that can be served simultaneously; is the current queuing rate, reflecting the busyness of the charging station; W j (t) is the charging station c j The average waiting time of the vehicle is used to further evaluate the waiting cost of the vehicle at the charging station; γ 1 It is an adjustment coefficient used to balance the impact of queuing rate and waiting time.

[0134] Time-of-use electricity price j : Reflects the charging cost at the current moment. The electricity price changes over time.

[0135] Charging service fee F j : Indicates additional charging service charges, usually a fixed fee or charged by service type.

[0136] The connection between the electric vehicle node and the charging station node is defined as a link (eij ), which is used to represent possible charging behaviors. The weight of a link is determined by the similarity between nodes and their internal attributes. The calculation of link weight combines historical encounter records and node internal information to more accurately reflect possible future charging behaviors.

[0137] In order to predict the possible links between electric vehicles and charging stations, the similarity calculation is based on the improved CAR-basedAdamic / Adar Index (CAA) indicator combined with the time decay function. The specific similarity calculation formula is as follows:

[0138]

[0139] Among them, S(x,y) is the similarity score between node x and node y, Γ(x) is the set of neighbor nodes of node x, Γ(y) is the set of neighbor nodes of node y, k(i) is the degree value of common neighbor node i, γ(i) is the subset of node i’s neighbors, p(t,t i,x ,t i,y ) is the time decay function, λ is the time decay coefficient, t is the current time slot, t i,x is the time slot when neighbor node i and node x last met, t i,y is the time slot when neighbor node i and node y last met.

[0140] On the basis of basic similarity calculation, node internal information such as load forecast, time-of-use electricity price, charging service fee and passenger demand forecast is integrated into the similarity calculation to form a comprehensive similarity formula:

[0141]

[0142] Among them, S′(v i ,c j ) is the tram v i and charging station j The comprehensive similarity score between i ,c j ) is the tram v i and charging station j The similarity value between them, α, β, γ, δ are adjustment coefficients, L j For charging station c j The current load forecast value, P j For charging station c j The current time-of-use electricity price, F j For charging service fee, D i To meet the needs of tram users.

[0143] In this embodiment, the recommended target is to minimize the additional moving distance recommended by the charging station, so that the electric vehicle can minimize deviation from the originally planned driving path when charging. The formula for the recommended target is as follows:

[0144]

[0145] in, is the additional travel distance of the recommended charging station, ||c j -p i (t)|| is the tram v i From the current position p i (t) Go to charging station c j The distance, ||p i (t+τ)-c j || After charging from the charging station c j Move to the future destination p i (t+τ) distance, ||p i (t+τ)-p i (t)|| is the shortest path length from the current time t to the future time t+τ.

[0146] To improve the recommendation effect, the Top-K recommendation method is used to sort the similarity scores and select the K charging stations with the highest scores as the final recommendation results. The specific steps are as follows: Calculate the similarity matrix A x , where each element represents the similarity score between nodes; sort the results in the similarity matrix from large to small according to the score; select the first K charging stations with the highest scores to form a recommendation sequence Ω x (t)=(ω 1 ,ω 2 ,…,ω K ).

[0147] The above embodiments further illustrate the purpose, technical solutions and advantages of the present invention in detail. It should be understood that the above embodiments are only preferred implementation modes of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made to the present invention within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting a short-term charging route for electric vehicles based on environmental factors, characterized in that: include: Construct a traffic network model; Collect traffic flow data and denoise it using the exponential moving average method; The Spearman correlation coefficient formula was used to calculate the correlation of the AC flow data after denoising; The traffic flow data is normalized according to the correlation, and the normalized data is input into the BP-LSTM-Attention prediction model to obtain the traffic flow in the next time period; the flow-density-speed relationship is constructed according to the predicted traffic flow; the road impedance model is constructed according to the flow-density-speed relationship; the travel time path of the tram is obtained according to the road impedance model, and the mileage power consumption of the tram is calculated; the mileage power consumption is processed by the Monte Carlo method to obtain the prediction result of the spatiotemporal distribution of the tram charging load; a tram link model is constructed, and the tram link model consists of multiple charging pile nodes; the optimized similarity calculation formula is used to calculate the similarity between the charging pile node and the tram node, and the link between the two nodes is determined according to the similarity; the comprehensive similarity formula is used to calculate the comprehensive similarity of the link according to the prediction results of the spatiotemporal distribution of the link and the tram charging load; the optimal driving route from the tram node to the charging pile node is recommended according to the comprehensive similarity.

2. A method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: Constructing a traffic network model includes using graph theory to model the traffic network, and its expression is: Among them, G represents the traffic road network of the mountain city; J represents the set of all nodes in the graph G; E represents the set of all road sections in the graph G; W is the set of road section weights, that is, road resistance; K represents the time period set, dividing the whole day into m time periods; Q is the traffic volume on the road; V is the vehicle speed on the road; θ is the road section slope of the mountain city road; n represents the number of nodes.

3. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The BP-LSTM-Attention prediction model is used to process the input data, including: screening the collected relevant traffic flow data, and constructing the weather factor traffic flow time series matrix according to the time series; normalizing the weather factor traffic flow time series matrix to obtain a standard data matrix; sampling the standard data matrix using the sliding window method to obtain a time series; preprocessing the time series; inputting the preprocessed sequence into the LSTM layer to obtain the dynamic features of the time series; inputting the dynamic features of the time series into the hidden layer for feature enhancement; inputting all enhanced dynamic features into the Attention layer for weight fusion to obtain the fused dynamic features; inputting the fused dynamic features into the classification prediction layer to obtain the classification prediction results.

4. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: Constructing the flow-density-speed relationship includes: The speed expression is: V=V max ×e -αθ The relationship between velocity, flow and density is expressed as: Q=VK The density expression is: The relationship between velocity and density is: Among them, α is the adjustment coefficient, V max is the maximum speed limit of mountainous urban roads; L is the length of the road section; K represents the vehicle density; N is the number of vehicles in the road section; V f represents the free flow speed; K j represents the blocking density.

5. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The shortest travel time path of the tram is obtained by using the improved Floyd algorithm to process the data collected from each node to obtain the shortest travel time path; The specific steps include: Step 1: Initialize the adjacency matrix w (0) , set the initial time cost between i and j to If points i and j are not directly connected, then T total =∞; Step 2: Construct w (1) , insert the first intermediate node between i and j to optimize the time cost, calculate Step 3: Construct w (2) , insert a second node between the two points and calculate Step 4: Construct w (n) ,calculate Then (n) represents the shortest time cost between i and j after traversing all nodes, w (n) It forms the shortest time cost matrix between each node.

6. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: Calculating the electric power consumption of a tram includes: Among them, E d is the power consumption of the tram's mileage, represents the power consumption per unit mileage of EV when driving on a k-level road, m is the driving path from the starting point to the destination, and x i,j is the horizontal distance between points i and j, a i,j is the climbing coefficient a c ,h i,j is the relative height difference between points i and j in the vertical direction, ε k is the power consumption per unit mileage of EV when driving on a k-level road, m T is the proportional coefficient of energy consumption change at different temperatures.

7. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The Monte Carlo method is used to process the mileage power consumption, including: constructing the probability distribution of user travel behavior in each time period in the future through the Monte Carlo method, generating a variety of potential charging scenarios through extensive random simulation operations, summarizing all charging scenarios, and obtaining the probability distribution of load; planning the user's driving path according to the load probability distribution, using the Floyd algorithm to plan the user's travel path with the shortest driving time as the goal, obtaining the travel time of each road based on the speed-flow-density model, and updating its spatiotemporal variables and charge state after the EV arrives at the destination, and calculating the travel power consumption based on the duration and mileage of the trip, combined with the impact of terrain and temperature on energy consumption; updating the remaining power according to the travel power consumption; analyzing the user's charging needs, obtaining the driving spatiotemporal distribution of each electric vehicle and the information at each destination, when the vehicle arrives at a destination, judging whether the user needs to charge based on the remaining power, if charging is required, determining the charging time based on the residence time and power situation, calculating the charging load of the electric vehicle at each destination, i.e., each functional area, and continuing to drive if charging is not required.

8. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The optimized similarity calculation formula is used to calculate the similarity between the charging pile node and the tram node, including: Among them, S(x, y) is the similarity score between node x and node y, Γ(x) is the set of neighbor nodes of node x, Γ(y) is the set of neighbor nodes of node y, k(i) is the degree value of common neighbor node i, γ(i) is the subset of node i’s neighbors, p(t, t i，x , t i，y ) is the time decay function, λ is the time decay coefficient, t is the current time slot, t i，x is the time slot when neighbor node i and node x last met, t i，y is the time slot when neighbor node i and node y last met.

9. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The comprehensive similarity formula is: Among them, S′(v i , c j ) is the tram v i and charging station j The comprehensive similarity score between i , c j ) is the tram v i and charging station j The similarity value between them, θ, β, γ, δ are adjustment coefficients, L j For charging station c j The current load forecast value, P j For charging station c j The current time-of-use electricity price, F j For charging service fee, D i To meet the needs of tram users.

10. The method for predicting a short-term charging route of an electric vehicle based on environmental factors according to claim 1, characterized in that: The formula for recommending the optimal driving route from the tram node to the charging pile node is: in, is the additional travel distance of the recommended charging station, ||c j -p i (t)|| is the tram v i From the current position p i (t) Go to charging station c j The distance, ||p i (t+τ)-c j || After charging from the charging station c j Move to the future destination p i (t+τ) distance, ||p i (t+τ)-p i (t)|| is the shortest path length from the current time t to the future time t+τ.

Images