Electric vehicle load prediction method considering traffic road conditions

Abstract

The invention belongs to the technical field of novel electric power system electric vehicle loads, and particularly relates to an electric vehicle load prediction method considering traffic road conditions. The method comprises the following steps: analyzing travel behavior characteristics of an electric vehicle user and modeling; performing clustering analysis and modeling on the user travel feature model; predicting the future electric vehicle ownership to obtain future electric vehicle ownership growth condition prediction data; and performing large-scale electric vehicle charging load prediction on the obtained future electric vehicle ownership increase condition prediction data to obtain a future electric vehicle load prediction result considering the traffic road condition. According to the method, the traffic road conditions encountered when the electric vehicle user travels are clustered, and the future electric vehicle ownership is predicted by using the artificial intelligence algorithm, so that the result of electric vehicle load prediction performed by using the Monte Carlo method subsequently is more accurate than the traditional electric vehicle charging load prediction result; the value is closer to a real value.

Description

Technical Field

[0001] The invention belongs to the technical field of electric vehicle load in a new power system, and in particular relates to an electric vehicle load prediction method taking traffic road conditions into consideration. Background Art

[0002] Combined with the national requirements and plans for the promotion and application of new energy vehicles, the number of electric vehicles in the country has increased year by year, and now exceeds 5 million vehicles per year. Therefore, vigorously developing electric vehicles and gradually replacing traditional fuel vehicles has gradually become a future development trend.

[0003] As the penetration rate of electric vehicles continues to increase, large-scale electric vehicles will be connected to the power grid in the future, and their charging and discharging behaviors will bring challenges to the stable operation of the power grid. Accurately predicting the charging demand of electric vehicles is the basis for studying the two-way interaction of electric vehicles on the power grid, and is of great significance for analyzing the impact of electric vehicles connected to the power grid, power system planning and stable operation.

[0004] The demand for electric vehicle charging is affected by many factors, and the charging load of electric vehicles has strong randomness and volatility, which makes it difficult to predict the charging demand. In recent years, domestic and foreign scholars have done a lot of relevant research on the prediction of electric vehicle charging load. There is a load prediction model established based on the statistical probability distribution of travel characteristic data, calculating the total charging demand of a large number of electric vehicles under different penetration rates, considering the charging demand of multiple types of electric vehicles, and calculating the charging demand through charging influencing factors such as the initial state of charge, daily mileage, and charging time, but the selection of charging time distribution is somewhat subjective. There are also settings for users to charge immediately when they return home or to charge when the charge is lower than a certain threshold, but all of them consider that users only use home charging piles and ignore the calculation of the charging load of public charging piles.

[0005] Some methods use big data related technologies to establish prediction models through historical traffic and weather data, and use decision trees to establish classification standards to predict the charging demand of electric vehicles on different date types. However, the above methods do not consider the factors affecting charging comprehensively enough.

[0006] The existing charging demand prediction methods at home and abroad only consider conventional influencing factors such as the number of electric vehicles, daily mileage, and initial state of charge when modeling, but ignore the impact of road conditions on the speed of electric vehicle charge consumption, resulting in deviations in the prediction results. Therefore, further research and development and breakthroughs are urgently needed by those skilled in the art. Summary of the invention

[0007] In view of the shortcomings of the above-mentioned prior art methods for predicting electric vehicle charging loads, which tend to ignore the impact of traffic road conditions on the electric vehicle state of charge SOC and the speed of electric vehicle charge consumption when modeling, the present invention provides an electric vehicle load prediction method that takes traffic road conditions into consideration. Its purpose is to consider the impact of traffic road conditions on driving speed and state of charge, combine the charging time and state of charge probabilistic model obtained from statistical data, use the K-medoid clustering method to cluster the traffic road conditions encountered by electric vehicle users, classify and establish a more accurate load prediction model, calculate the electric vehicle charging demand through the Monte Carlo method to obtain the daily charging load prediction curve, and analyze the calculation results, so as to achieve the invention purpose of more accurate electric vehicle charging load prediction results.

[0008] The technical solution adopted by the present invention to achieve the above-mentioned purpose is:

[0009] A method for predicting electric vehicle load considering traffic road conditions comprises the following steps:

[0010] Analyze and model the travel behavior characteristics of electric vehicle users;

[0011] Conduct cluster analysis and model the user travel characteristics model;

[0012] Predict the number of electric vehicles in the future and obtain the forecast data on the growth of the number of electric vehicles in the future;

[0013] Based on the forecast data of future electric vehicle ownership growth, large-scale electric vehicle charging load forecasting is carried out to obtain the electric vehicle load forecast results taking into account traffic road conditions.

[0014] Furthermore, the travel behavior characteristics of electric vehicle users are analyzed and modeled, including: probability distribution of starting charging time, probability distribution of daily mileage, and probability distribution of battery capacity.

[0015] Furthermore, the probability distribution of the starting charging time is as follows:

[0016]

[0017] In the above formula, x represents the time, and the mean μ t =17.6, variance σ t =3.4, and plotted using MATLAB, we can get the probability distribution of the charging start time of ordinary special electric vehicles;

[0018] The daily mileage probability distribution is shown as follows:

[0019]

[0020] In the above formula, the mean μ s =3.3, variance σ s = 0.88, and the daily mileage distribution of ordinary private electric vehicles is obtained by plotting with MATLAB;

[0021] The battery capacity probability distribution is shown as follows:

[0022]

[0023] In the above formula, f C is the probability distribution density function of different electric vehicle battery capacities.

[0024] Furthermore, the user travel feature model is clustered and modeled, and the K-medoid clustering algorithm is used to divide the traffic conditions into m categories by clustering. The amount of charge to be charged in each category is Ci, and the overall amount of charge to be charged in the area is calculated as shown in formula (4). The overall charge amount is the accumulation of the charge amount of electric vehicles in each cluster category:

[0025]

[0026] In the above formula, C tot is the total charging capacity, n is the number of electric vehicles;

[0027] According to the clustering results, three traffic conditions, namely congested, normal and unobstructed, are modeled and analyzed respectively. The calculation method of the remaining overall SOC of each type of traffic condition at the starting charging time is shown in formula (5):

[0028]

[0029] In the above formula, SOC start is the state of charge during this charging; β i The weight of each category; SOC end It is the state of charge after the last charge and before driving; L c is the daily mileage; α η is the influencing factor of congestion; Z is the power consumption per kilometer of electric driving; L is the cruising range of the electric vehicle. Different road congestion conditions directly affect the power consumption of the electric vehicle per 100 kilometers; n is the number of electric vehicles in each category after clustering; the charge to be charged for each type of road condition is calculated and added to the overall charge amount. The charge state of the electric vehicle is negatively correlated with the mileage. Road congestion will result in higher power consumption for the same mileage.

[0030] Furthermore, the prediction of the future number of electric vehicles and the prediction data of the future growth of the number of electric vehicles are obtained by using the BP neural network algorithm. 20 neurons are selected during training, 70% of the original data are selected as the training group, 15% of the original data are selected as the verification group, and 15% of the original data are selected as the test group. Sigmoid differentiable functions and linear functions are used as the excitation functions of the neural network, the S-type tansig tangent function is selected as the hidden layer excitation function, and the S-type tansig logarithmic function is selected as the output layer function.

[0031] Furthermore, the prediction of the future electric vehicle ownership growth forecast data is used to predict the charging load of electric vehicles on a large scale, and the prediction result of the electric vehicle load considering the traffic road conditions is obtained by extracting the charging load influencing factor samples through the Monte Carlo method, calculating the charging load of electric vehicles, and performing cluster analysis on the two-dimensional plane formed by the driving time and driving distance characteristics of each vehicle on the basis of obtaining the electric vehicle ownership and establishing the probability model of the starting charging time and the starting state of charge. The traffic road condition index is divided into three levels of congestion, normal and smooth through clustering, and the state of charge distribution parameters of the two traffic levels are respectively obtained to establish the congestion factor influencing factor model, and then the state of charge model of the electric vehicle before charging is quantitatively established. The daily charging load of each electric vehicle is calculated by repeatedly sampling the starting charging time and the starting state of charge, and then the overall charging demand of the two types of vehicles, private cars and electric buses, is calculated;

[0032] The whole day is divided into 24 computing nodes, and the charging load is calculated once every hour. The charging power calculation is shown in formula (6):

[0033]

[0034] In the above formula, P is the total charging power, and Pi is the charging power of the vehicle in the i-th hour.

[0035] An electric vehicle load prediction device considering traffic road conditions, comprising:

[0036] Travel behavior characteristics analysis and modeling module, used to analyze and model the travel behavior characteristics of electric vehicle users;

[0037] Cluster analysis and modeling module, used to perform cluster analysis and modeling on user travel feature models;

[0038] The ownership prediction module is used to predict the future ownership of electric vehicles and obtain the forecast data of the future growth of the ownership of electric vehicles;

[0039] The charging load prediction module is used to perform large-scale electric vehicle charging load prediction based on the future electric vehicle ownership growth forecast data, and obtain the electric vehicle load prediction result taking into account the traffic road conditions.

[0040] Furthermore, the electric vehicle load prediction device taking traffic road conditions into consideration is used to implement any of the steps of the electric vehicle load prediction method taking traffic road conditions into consideration.

[0041] A computer device comprises a storage medium, a processor and a computer program stored on the storage medium and executable on the processor, wherein when the processor executes the computer program, the steps of any one of the methods for predicting electric vehicle load taking into account traffic road conditions are implemented.

[0042] A computer storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of any one of the methods for predicting electric vehicle load taking into account traffic road conditions are implemented.

[0043] The present invention has the following beneficial effects and advantages:

[0044] The present invention proposes an electric vehicle load prediction method taking into account traffic road conditions, which makes up for the fact that the traditional Monte Carlo method-based electric vehicle load prediction ignores the impact of road congestion on the power consumption per unit mileage of the electric vehicle. The traditional Monte Carlo method simply uses a fixed value of the linear relationship between mileage and power consumption to calculate the electric vehicle load. The present invention makes the electric vehicle load prediction closer to the actual value.

[0045] The present invention aims to improve the accuracy of charging load prediction of electric vehicles in a typical day. To this end, the present invention takes into account the impact of traffic conditions on driving speed and state of charge. First, the travel characteristics of electric vehicle users are modeled. Then, the K-medoid clustering method is used to cluster the traffic road conditions encountered by electric vehicle users in combination with the charging time and state of charge probabilistic model obtained from statistical data. A more accurate load prediction model is established by classification. Then, the BP neural network algorithm is used to predict the future number of electric vehicles. Finally, the Monte Carlo method is used to calculate the charging demand of electric vehicles to obtain a daily charging load prediction curve, and the calculation results are analyzed.

[0046] The present invention performs cluster analysis on user travel characteristics to determine the traffic congestion of each user when traveling, and then sets the power consumption per kilometer for the clustered electric vehicle users, so that the subsequent Monte Carlo method for electric vehicle load prediction results are more in line with the actual value. The present invention also predicts the future number of electric vehicles in a certain area through an artificial intelligence algorithm, so that the subsequent Monte Carlo method for electric vehicle load prediction results are more accurate than the traditional electric vehicle charging load prediction results.

[0047] The present invention takes into account the traffic road conditions when electric vehicle users travel and uses the K-medoid clustering method to cluster the traffic road conditions encountered by electric vehicle users, and adopts the BP neural network to predict the future electric vehicle ownership. Compared with the traditional Monte Carlo simulation of electric vehicle load based only on the travel behavior of electric vehicle users, it is more in line with reality and greatly improves the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0049] Figure 1 It is a map of electric vehicle ownership in a certain region from 2014 to 2023;

[0050] Figure 2 is the probability distribution of the time when the electric vehicle starts charging according to the present invention;

[0051] Figure 3 is a probability distribution diagram of daily mileage of an electric vehicle of the present invention;

[0052] Figure 4 It is the travel time-mileage distribution diagram of the present invention;

[0053] Figure 5 is a traffic road condition clustering result diagram of the present invention;

[0054] Figure 6 This is a graph showing the prediction results of the number of electric vehicles in use using the BP neural network of the present invention. DETAILED DESCRIPTION

[0057] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0058] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0059] Refer to the following Figure 1-Figure 6 The technical solutions of some embodiments of the present invention are described.

[0060] Example 1

[0061] The present invention provides an embodiment, which is a method for predicting electric vehicle load taking into account traffic road conditions. Figure 1 As shown, Figure 1 It is a structural schematic diagram of the present invention.

[0062] The present invention provides an electric vehicle load prediction method considering traffic road conditions, comprising the following steps:

[0063] Step 1. Analyze and model the travel behavior characteristics of electric vehicle users.

[0064] The modeling of electric vehicle user travel behavior characteristics includes: the probability distribution of the starting charging time, the probability distribution of daily mileage, and the probability distribution of battery capacity.

[0065] The start charging time of a vehicle is affected by the type of vehicle, season, whether it is a holiday, and the user's personal habits. Through the investigation and analysis of urban traffic data, the start charging time is quantitatively analyzed using a known probability model. According to statistical data analysis and image observation, the data is more in line with the normal distribution. The present invention takes the charging model of an ordinary private electric vehicle as an example, and the probability distribution of the start charging time is shown as follows:

[0066]

[0067] In the above formula, x represents the time, and the mean μ t =17.6, variance σ t =3.4, and by plotting with MATLAB, we can get the probability distribution of the charging start time of ordinary special electric vehicles.

[0068] Considering that electric vehicles are a relatively new means of transportation and their sample size is still small, the present invention assumes that electric vehicle users have the same car habits as ordinary fuel vehicle users, selects the most typical private electric vehicles as data representing the family car usage behavior pattern, and concludes that the daily mileage follows a log-normal distribution. The probability distribution of daily mileage is shown in the following formula:

[0069]

[0070] In the above formula, the mean μs =3.3, variance σ s = 0.88. Using MATLAB, we can get the daily mileage distribution of ordinary private electric vehicles.

[0071] For the convenience of analysis, the present invention selects a uniform distribution within the range of 30-40 kw·h to represent the battery capacity of different electric vehicles. The density function of the probability distribution of the battery capacity of different electric vehicles is shown as follows:

[0072]

[0073] In the above formula, f C is the probability distribution density function of different electric vehicle battery capacities.

[0074] Step 2. Based on the user travel characteristics, that is, the "model established by the electric vehicle user travel behavior obtained in step 1", a cluster analysis is performed on the traffic road conditions encountered by the electric vehicle users.

[0075] The present invention takes into account the power consumption characteristics under different road conditions, and establishes a load prediction model based on the characteristics of this situation. The two-dimensional plane composed of the two characteristics of the driving time and the driving distance of each journey is clustered and analyzed, and the traffic condition index is divided into two levels of congestion and smoothness through clustering. The state of charge distribution parameters of the two traffic levels are obtained respectively, and a congestion factor influencing factor model is established, and then the state of charge model of the electric vehicle before charging is quantitatively established. The K-medoid clustering algorithm has a good clustering effect for normally distributed clusters and has good scalability when processing large data sets. Therefore, the present invention adopts the K-medoid clustering algorithm. Assuming that the traffic conditions are divided into m categories by clustering, Ci is the amount of charge to be charged in each category, then the overall amount of charge to be charged in the area is calculated as shown in formula (4), and the overall charging amount is the accumulation of the charging amount of electric vehicles in each clustering category.

[0076]

[0077] In the above formula, C tot is the total charging capacity, and n is the number of electric vehicles.

[0078] According to the clustering results, the three traffic conditions of congestion, normal and smooth are modeled and analyzed respectively. The calculation method of the remaining overall SOC of each type of traffic condition at the starting charging time is shown in formula (5):

[0079]

[0080] In the above formula, SOC start is the state of charge during this charging; β i The weight of each category; SOC endIt is the state of charge after the last charge and before driving; L c is the daily mileage; α η is the influencing factor of congestion; Z is the power consumption per kilometer of electric driving; L is the cruising range of electric vehicles. Different road congestion conditions directly affect the power consumption of electric vehicles per 100 kilometers; n is the number of electric vehicles in each category after clustering. The amount of charge to be charged for each type of road condition is calculated and added to the overall charge amount. The state of charge of electric vehicles is negatively correlated with the mileage. Road congestion will cause higher power consumption for the same mileage.

[0081] Step 3. Forecast the future number of electric vehicles and obtain the forecast data of the future growth of the number of electric vehicles, so as to predict the future electric vehicle load.

[0082] The principle of BP neural network is to continuously learn based on the training samples of known input and output vectors, and adjust and correct the weights and thresholds between neurons, so that the network continuously approaches the mapping relationship between sample input and output, and then uses test samples to detect the accuracy of the results. The BP neural network model is used to train the data samples of the number of electric vehicles in a certain area in the past to predict the growth of the number of electric vehicles in a certain area in the next seven years.

[0083] When the BP neural network is used to train the electric vehicle ownership data, the number of neurons selected during training is 20, 70% of the original data is selected as the training group, 15% of the original data is selected as the verification group, and 15% of the original data is selected as the test group. The present invention uses Sigmoid differentiable functions and linear functions as the excitation functions of the neural network, selects the S-type tansig tangent function as the hidden layer excitation function, and selects the S-type tansig logarithmic function as the output layer function. The parameter settings of the BP neural network ownership prediction model are shown in Table 1.

[0084] Table 1 Parameters of BP neural network inventory prediction model

[0085]

[0086] Step 4. Based on the obtained forecast data of future electric vehicle ownership growth, a large-scale electric vehicle charging load forecast is conducted to obtain the electric vehicle load forecast result taking into account the traffic road conditions.

[0087] The future number of electric vehicles obtained in step 3 is used as the basic data for predicting the scale of electric vehicles in step 4.

[0088] The typical method of describing the charging load of electric vehicles by probability statistics is to extract samples of factors affecting the charging load through the Monte Carlo method. To calculate the charging load of electric vehicles, based on the number of electric vehicles, the probability model of the starting charging time and the starting state of charge is established, and then a cluster analysis is performed on the two-dimensional plane composed of the two characteristics of each vehicle's driving time and driving distance. The traffic condition index is divided into three levels: congestion, normal and smooth through clustering. The state of charge distribution parameters of the two traffic levels are obtained respectively, and the congestion factor influencing factor model is established. Then, the state of charge model before charging of electric vehicles is quantitatively established. Then, the daily charging load of each electric vehicle is calculated by repeatedly sampling the starting charging time and the starting state of charge, and then the overall charging demand of private cars and electric buses is calculated. The whole day is divided into 24 computing nodes, and the charging load is calculated once every 1 hour. The charging power is calculated as shown in formula (6), where P is the total charging power and Pi is the charging power of the vehicle in the i-th hour.

[0089]

[0090] Example 2

[0091] The present invention provides an embodiment, which is an electric vehicle load prediction method considering traffic road conditions. Different from the first embodiment, this embodiment takes the use of electric vehicles in a certain place in my country as an example and takes private electric vehicles as the main research object. Figure 1 The figure shows the survey results of electric vehicle ownership in a certain place in my country over the past ten years, which will serve as the basis for the subsequent prediction of future electric vehicle ownership and the calculation of electric vehicle load using the Monte Carlo method.

[0092] As described in Example 1, step 1 analyzes and models the travel behavior of electric vehicle users, and the specific implementation method is as follows:

[0093] According to the probability distribution of the starting charging time of private electric vehicles and the probability mathematical model of the daily mileage of electric vehicles given in the invention content, the function is plotted using MATLAB, as shown in FIG. Figure 2 as well as Figure 3 , which serves as the basis for large-scale electric vehicle charging load forecasting in step 4.

[0094] The specific implementation method of performing cluster analysis and modeling on the user travel feature model in step 2 of Example 1 is as follows:

[0095] The present invention takes into account the power consumption characteristics under different road conditions and establishes a load prediction model based on the characteristics of this situation. Figure 4As shown in the figure, the K-medoid clustering algorithm is used to perform cluster analysis on the driving time-mileage two-dimensional plane class, and the traffic road conditions encountered by each electric vehicle user are divided into three categories: smooth, normal and congested. The clustering results are shown in Figure 5 As shown, the power consumption per kilometer of electric vehicles varies depending on the traffic road conditions.

[0096] For the vehicle models considered in the present invention, the road conditions are divided into three types: congested, normal, and unobstructed. When the road condition is unobstructed, the electric vehicle is relatively energy-saving, with a power consumption of 13.8 kWh per 100 kilometers; under normal conditions, the electric vehicle consumes 15.2 kWh per 100 kilometers, and under congested conditions, the electric vehicle consumes 20.7 kWh per 100 kilometers. These three data are also the basis for the large-scale electric vehicle charging load prediction in step 4.

[0097] The specific implementation method of predicting the number of electric vehicles in use as described in step 3 of Example 1 to obtain the forecast data of the future growth of the number of electric vehicles in use is as follows:

[0098] In order to better fit the actual situation, the present invention is based on the number of electric vehicles in a certain county in my country in the past ten years. Figure 1 As shown, the future electric vehicle ownership is predicted as the data basis for step 4. The specific parameter settings of the BP neural network are as described in Example 1. The original data are trained using the BP neural network algorithm. The specific parameter settings of the BP neural network are shown in Table 1. 70% of the original data are selected as the training group, 15% of the original data are selected as the verification group, and 15% of the original data are selected as the test group to predict the future electric vehicle ownership of the county. The prediction results are as follows Figure 6 shown.

[0099] As described in step 4 of Example 1, the specific implementation method of performing large-scale electric vehicle charging load forecasting on the obtained electric vehicle ownership growth forecast data to obtain the electric vehicle load forecasting result considering the traffic road conditions is as follows:

[0100] An electric vehicle load prediction model considering traffic road conditions constructed by the present invention takes private electric vehicles in a certain county in my country as the research object. With the above three steps as the foundation, the number of electric vehicles is predicted, the travel characteristics of electric vehicle users are modeled, and the traffic road conditions encountered by electric vehicle users are clustered and divided into three groups. The power consumption of electric vehicles per 100 kilometers under different traffic road conditions is also different. Then, the three types of electric vehicles are subjected to Monte Carlo simulation respectively, and the charging time, daily mileage and starting SOC of the electric vehicles are randomly selected. The electric vehicle users of the three groups are simulated respectively, and finally the total load of electric charging is obtained by superimposing them together, and the charging load of electric vehicles in a typical day under the two conditions of considering traffic road conditions and not considering traffic is obtained.

[0101] In summary, it can be seen that when considering the traffic road conditions encountered by electric vehicle users, the peak value of the electric vehicle charging load will be higher than the load when the traffic road conditions are not considered, and in some time periods it will be lower than the load when the traffic road conditions are not considered, which is closer to the actual value. The experimental results further demonstrate the accuracy and superiority of the present invention in predicting electric vehicle load.

[0102] Example 3

[0103] The present invention further provides an embodiment, which is an electric vehicle load prediction device taking into account traffic road conditions, comprising:

[0104] Travel behavior characteristics analysis and modeling module, used to analyze and model the travel behavior characteristics of electric vehicle users;

[0105] Cluster analysis and modeling module, used to perform cluster analysis and modeling on user travel feature models;

[0106] The ownership prediction module is used to predict the future ownership of electric vehicles and obtain the forecast data of the future growth of the ownership of electric vehicles;

[0107] The charging load prediction module is used to perform large-scale electric vehicle charging load prediction based on the future electric vehicle ownership growth forecast data, and obtain the electric vehicle load prediction result taking into account the traffic road conditions.

[0108] The electric vehicle load prediction device considering traffic road conditions described in this embodiment is used to implement the steps of the electric vehicle load prediction method considering traffic road conditions described in any one of Embodiments 1 or 2.

[0109] Example 4

[0110] Based on the same inventive concept, an embodiment of the present invention further provides a computer device, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor. When the processor executes the computer program, the steps of any one of the electric vehicle load prediction methods considering traffic road conditions described in Embodiment 1 or 2 are implemented.

[0111] Example 5

[0112] Based on the same inventive concept, an embodiment of the present invention further provides a computer storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any one of the electric vehicle load forecasting methods considering traffic road conditions described in Embodiment 1 or 2 are implemented.

[0113] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0114] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for predicting electric vehicle load considering traffic road conditions, characterized by: The following steps are involved: Analyze and model the travel behavior characteristics of electric vehicle users; Conduct cluster analysis and model the user travel characteristics model; Predict the number of electric vehicles in the future and obtain the forecast data on the growth of the number of electric vehicles in the future; Based on the forecast data of future electric vehicle ownership growth, large-scale electric vehicle charging load forecasting is carried out to obtain the electric vehicle load forecast results taking into account traffic road conditions.

2. The electric vehicle load forecasting method considering traffic road conditions according to claim 1 is characterized by: The travel behavior characteristics of electric vehicle users are analyzed and modeled, including: probability distribution of starting charging time, probability distribution of daily mileage, and probability distribution of battery capacity.

3. The electric vehicle load forecasting method considering traffic road conditions according to claim 2 is characterized by: The probability distribution of the starting charging time is shown as follows: In the above formula, x represents the time, and the mean μ t =17.6, variance σ t =3.4, and plotted using MATLAB, we can get the probability distribution of the charging start time of ordinary special electric vehicles; The daily mileage probability distribution is shown as follows: In the above formula, the mean μ s =3.3, variance σ s = 0.88, and the daily mileage distribution of ordinary private electric vehicles is obtained by plotting with MATLAB; The battery capacity probability distribution is shown as follows: In the above formula, f C is the probability distribution density function of different electric vehicle battery capacities.

4. The electric vehicle load forecasting method considering traffic road conditions according to claim 1 is characterized by: The user travel feature model is clustered and analyzed and modeled. The K-medoid clustering algorithm is used to divide the traffic conditions into m categories by clustering. The amount of charge to be charged in each category is Ci. The overall amount of charge to be charged in the area is calculated as shown in formula (4). The overall charge amount is the accumulation of the charge amount of electric vehicles in each cluster category: In the above formula, C tot is the total charging capacity, n is the number of electric vehicles; According to the clustering results, three traffic conditions, namely congested, normal and unobstructed, are modeled and analyzed respectively. The calculation method of the remaining overall SOC of each type of traffic condition at the starting charging time is shown in formula (5): In the above formula, SOC start is the state of charge during this charging; β i The weight of each category; SOC end It is the state of charge after the last charge and before driving; L c is the daily mileage; α η is the influencing factor of congestion; Z is the power consumption per kilometer of electric driving; L is the cruising range of the electric vehicle. Different road congestion conditions directly affect the power consumption of the electric vehicle per 100 kilometers; n is the number of electric vehicles in each category after clustering; the charge to be charged for each type of road condition is calculated and added to the overall charge amount. The charge state of the electric vehicle is negatively correlated with the mileage. Road congestion will result in higher power consumption for the same mileage.

5. The electric vehicle load forecasting method considering traffic road conditions according to claim 1 is characterized by: The prediction of the future electric vehicle ownership and the acquisition of the predicted data on the growth of the future electric vehicle ownership are performed by using the BP neural network algorithm. 20 neurons are selected during training, 70% of the original data are selected as the training group, 15% of the original data are selected as the verification group, and 15% of the original data are selected as the test group. Sigmoid differentiable functions and linear functions are used as the excitation functions of the neural network, the S-type tansig tangent function is selected as the hidden layer excitation function, and the S-type tansig logarithmic function is selected as the output layer function.

6. The electric vehicle load forecasting method considering traffic road conditions according to claim 1 is characterized by: The method of performing large-scale electric vehicle charging load prediction on the obtained future electric vehicle ownership growth forecast data to obtain the electric vehicle load prediction result considering the traffic road conditions is to extract charging load influencing factor samples through the Monte Carlo method, calculate the electric vehicle charging load, obtain the electric vehicle ownership, establish the starting charging time and the starting state of charge probability model, perform cluster analysis on the two-dimensional plane formed by the driving time and driving distance characteristics of each vehicle, divide the traffic road condition index into three levels of congestion, normal and smooth through clustering, respectively obtain the state of charge distribution parameters of the two traffic levels, establish the congestion factor influencing factor model, and then quantitatively establish the electric vehicle state of charge model before charging, calculate the daily charging load of each electric vehicle by repeatedly sampling the starting charging time and the starting state of charge, and then calculate the overall charging demand of the two types of vehicles, namely private cars and electric buses; The whole day is divided into 24 computing nodes, and the charging load is calculated once every hour. The charging power calculation is shown in formula (6): In the above formula, P is the total charging power, and Pi is the charging power of the vehicle in the i-th hour.

7. An electric vehicle load prediction device taking into account traffic road conditions, characterized by: include: Travel behavior characteristics analysis and modeling module, used to analyze and model the travel behavior characteristics of electric vehicle users; Cluster analysis and modeling module, used to perform cluster analysis and modeling on user travel feature models; The ownership prediction module is used to predict the future ownership of electric vehicles and obtain the forecast data of the future growth of the ownership of electric vehicles; The charging load prediction module is used to perform large-scale electric vehicle charging load prediction based on the future electric vehicle ownership growth forecast data, and obtain the electric vehicle load prediction result taking into account the traffic road conditions.

8. The electric vehicle load prediction device considering traffic road conditions according to claim 7 is characterized in that: The device is used to implement the steps of an electric vehicle load prediction method taking into account traffic road conditions as described in any one of claims 2-6.

9. A computer device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the electric vehicle load forecasting method taking into account traffic road conditions as described in any one of claims 1 to 6 are implemented.

10. A computer storage medium, characterized in that: The computer storage medium stores a computer program, and when the computer program is executed by the processor, the steps of the electric vehicle load prediction method considering traffic road conditions described in any one of claims 1 to 6 are implemented.

Images