Method for constructing electric vehicle charging load prediction model considering multiple power consumption factors

Abstract

The application provides a kind of electric vehicle charging load prediction model construction method considering multiple power consumption factors, comprising determining the number of daily active electric vehicles;Analyzing the multiple power consumption factors of electric vehicles, and accordingly establishing a per-100-kilometer power consumption model to obtain the expected value of electric vehicle power consumption per 100 kilometers under consideration of all factors;Predict and generate a load curve representing the total charging demand of all electric vehicles in the target area over time. By comprehensively considering the multiple power consumption factors of electric vehicles, using Monte Carlo simulation method and probability distribution model, firstly, the accurate prediction of electric vehicle charging load is realized;Then by determining the number of daily active electric vehicles, the probability distribution of first charging time and driving distance is more accurately simulated, and finally by analyzing the air conditioner usage probability and the power consumption ratio in the on and off states, the power grid resource allocation is optimized, the peak load is reduced, and the stability and reliability of the power grid are improved.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle charging energy consumption identification and analysis technology, specifically to a method for constructing an electric vehicle charging load prediction model that considers multiple power consumption factors. Background Technology

[0002] Currently, the technology of new energy electric vehicles is developing rapidly, and their market share is constantly increasing. However, the charging of electric vehicles is highly random in terms of time and space. Large-scale grid connection will inevitably have adverse effects on grid operation, such as affecting grid stability and generating numerous harmonics that cause voltage distortion. Modeling the large-scale electric vehicle charging load is fundamental to studying its impact on the power grid. Researchers have already studied various factors that may affect electric vehicle load and established load forecasting models, but the factors considered in the forecasts are still not comprehensive.

[0003] In real life, the power consumption of electric vehicles per 100 kilometers is affected by a variety of factors, such as the brand of the electric vehicle, the degree of battery aging, the degree of motor aging, the power consumption of in-vehicle equipment, and road conditions. In addition, existing studies have not considered the interference of idle electric vehicles on model building.

[0004] Therefore, it is essential to propose a method for constructing an electric vehicle charging load prediction model that considers multiple power consumption factors. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a method for constructing an electric vehicle charging load prediction model that considers multiple power consumption factors. To achieve this objective, the present invention employs the following technical solution: a method for constructing an electric vehicle charging load prediction model that considers multiple power consumption factors, comprising:

[0006] By acquiring and cleaning the electric vehicle ownership and travel statistics in the target area, excluding vehicles that have not been used for a long time, and distinguishing travel patterns between holidays and non-holidays, the number of active electric vehicles per day can be determined.

[0007] This study analyzes multiple power consumption factors of electric vehicles, including brand, road conditions, aging, and in-vehicle equipment usage, assesses the impact of these factors on power consumption, divides the power consumption per 100 kilometers of electric vehicles of different brands into multiple intervals, calculates the proportion of each brand in each interval, and then establishes a total power consumption model per 100 kilometers to obtain the expected power consumption per 100 kilometers of electric vehicles considering all factors.

[0008] Based on the total power consumption model, the Monte Carlo simulation method is used to set input parameters that affect the charging load prediction. The probability distribution of the first charging time and driving mileage is simulated through a probability distribution model. Finally, the load curve representing the total charging demand of all electric vehicles in the target area over time is predicted and generated.

[0009] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0010] This invention comprehensively considers multiple power consumption factors of electric vehicles, such as brand, road conditions, aging, and the usage of in-vehicle equipment. Using the Monte Carlo simulation method and probability distribution model, it first achieves accurate prediction of electric vehicle charging load. Then, by acquiring and cleaning the electric vehicle ownership and travel statistics in the target area, excluding vehicles that have not been used for a long time, and distinguishing between holiday and non-holiday travel patterns, it can determine the number of active electric vehicles every day, thereby more accurately simulating the probability distribution of first charging time and driving mileage.

[0011] Furthermore, by analyzing the probability of air conditioner use and the power consumption ratio in the on and off states, this invention can assess the impact of air conditioner use on power consumption, further improving the prediction accuracy of the model. This not only provides a scientific basis for power grid load management and electric vehicle charging strategy formulation, but also optimizes power grid resource allocation, reduces peak load, and improves the stability and reliability of the power grid, thus providing important support for the construction of smart grids and the orderly charging of electric vehicles. Attached Figure Description

[0012] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0013] Figure 1 This is a schematic diagram of the charging load prediction model construction method proposed in one embodiment of the present invention;

[0014] Figure 2 This is a schematic diagram of the electric vehicle charging load calculation process based on the Monte Carlo method proposed in one embodiment of the present invention. Detailed Implementation

[0015] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0016] The present invention will be further described in detail below with reference to the accompanying drawings, but this is not intended to limit the scope of the invention.

[0017] As an understanding of the technical concept and implementation principle of the present invention, the main purpose of the present invention is to solve the problems of insufficient consideration of factors and insufficient prediction accuracy in the existing electric vehicle charging load prediction model.

[0018] like Figure 1 - Figure 2 As shown in the figure, as an embodiment of the present invention, the proposed method for constructing an electric vehicle charging load prediction model that considers multiple power consumption factors comprehensively analyzes multiple power consumption factors of electric vehicles, including vehicle brand, road conditions, aging degree, and in-vehicle equipment usage, to construct a new prediction model. Using the Monte Carlo simulation method and probability distribution model, combined with actual travel statistics and ambient temperature changes, the model accurately simulates the charging behavior and driving mileage of electric vehicles, providing a scientific basis for power grid load management, optimizing charging strategies, and reducing peak load on the power grid.

[0019] Including specific steps:

[0020] S1. By acquiring and cleaning the electric vehicle ownership and travel statistics in the target area, vehicles that have not been used for a long time are excluded, and travel patterns on holidays and non-holidays are distinguished, thereby determining the number of active electric vehicles per day.

[0021] S2. Analyze multiple power consumption factors of electric vehicles, including brand, road conditions, aging, and usage of in-vehicle equipment, assess the impact of these factors on power consumption, divide the power consumption per 100 kilometers of electric vehicles of different brands into multiple intervals, and calculate the proportion of each brand in each interval.

[0022] S3. Establish a model of total power consumption per 100 kilometers to obtain the expected power consumption of electric vehicles per 100 kilometers considering all factors.

[0023] S4. Based on the total power consumption model, the Monte Carlo simulation method is used to set the input parameters that affect the charging load prediction. The probability distribution model is used to simulate the probability distribution of the first charging time and driving mileage. Finally, the load curve representing the total charging demand of all electric vehicles in the target area over time is predicted and generated.

[0024] In one embodiment of the present invention, step S1 specifically includes the following steps:

[0025] S1-1. Obtain accurate household electric vehicle ownership and travel statistics for the target area through questionnaires, contacting relevant organizations or platforms. Example: Design a detailed questionnaire based on actual needs, covering key information about electric vehicles, and distribute it through multiple online and offline channels to collect extensive data. Simultaneously, proactively contact traffic management departments, vehicle registration agencies, and electric vehicle charging service providers to obtain official statistics through cooperation and data sharing. Then, utilize existing electric vehicle management platforms or applications to collect user travel and charging data, further enriching the data sources. Finally, set up a database that is regularly updated to reflect the latest changes in electric vehicle ownership and travel patterns.

[0026] S1-2. Perform data cleaning. The purpose of data cleaning is to remove duplicate, erroneous, and incomplete data, eliminate interference from idle electric vehicles on charge prediction data, and obtain the total number of active electric vehicles in the local area each day. The process is as follows: After excluding vehicles that have not been used for a long time, vehicles used within the last two weeks are defined as active electric vehicles. The total number of active electric vehicles is calculated based on the following formula using data obtained from surveys or relevant institutions: (1)

[0027] In the formula, The total number of active electric vehicles; This represents the total number of electric vehicles that are not used for an extended period. The number of electric vehicles in the target area; This is used as an active factor for electric vehicles; furthermore, since the number of vehicles traveling each day differs between holidays and non-holidays, the expected value of daily vehicle travel is calculated over a 14-day period, considering both holidays and non-holidays. , Differentiate between holiday and non-holiday travel patterns:

[0028] In the formula, This refers to the number of holiday days, 14- This refers to the number of days that are not public holidays.

[0029] In one embodiment of the present invention, in step S2, when evaluating the impact of multiple power consumption factors on power consumption, it is necessary to consider multiple power consumption factors, including different electric vehicle brands, road conditions, the aging level of the electric vehicle, and the usage of in-vehicle equipment, in order to analyze the impact of each power consumption factor on power consumption during the electric vehicle's operation. It should be noted that the power consumption per 100 kilometers varies among different electric vehicle brands. Currently, the power consumption per 100 kilometers for mainstream electric vehicles on the market is between 9 and 21 kWh. Therefore, the power consumption ranges for different brands of electric vehicles are divided as [9 kWh, 12 kWh], (12 kWh, 15 kWh], (15 kWh, 18 kWh], and (18 kWh, 21 kWh). The analysis and / or evaluation process is as follows:

[0030] The percentage of tram brands in each section is denoted as follows: , , and By taking the average power consumption for each range, we can statistically analyze the proportion of different brands of electric vehicles in each power consumption range. , , and Calculate the power consumption per 100 kilometers. The unit is kWh, used to assess the impact of vehicle type on power consumption: (4)

[0031] It should be noted that in reality, road conditions are closely related to the local road infrastructure and vehicle ownership. Generally, the more roads there are, the higher their quality and the better the road conditions. However, the more vehicles there are, the more vehicles are on the road, leading to worse road conditions, especially during peak commuting hours. Electric vehicles need to frequently start, accelerate, and decelerate when road conditions are poor or traffic jams occur, which increases power consumption. Therefore, introducing an average congestion factor can help mitigate this issue. The power consumption is adjusted based on traffic conditions to assess the impact of traffic conditions on power consumption (EC2). (5), where, , The average congestion distance per 100 kilometers for electric vehicles. This represents the average increase in electricity consumption during periods of congestion.

[0032] Furthermore, the aging degree of an electric vehicle includes the aging degree of the battery and the aging degree of the motor. Battery aging and motor aging will increase the energy consumption per 100 kilometers of the electric vehicle. Therefore, the aging coefficients of the electric vehicle's battery and motor can be considered. and The impact of aging on power consumption is quantified using EC3, measured in kWh. (6) In specific implementation, the aging coefficient and All data are local statistics.

[0033] Furthermore, since the usage of in-vehicle equipment is mainly related to ambient temperature, which affects the utilization rate of electric vehicle air conditioning, the ambient temperature is obtained from the day's meteorological temperature data, and the change of ambient temperature over time is denoted as a function. , Trigonometric functions were used for fitting, and the probability of the vehicle's air conditioning being turned on under different ambient temperatures was fitted using a Cauchy distribution. Based on this, the impact of air conditioning use on power consumption can be assessed by analyzing the probability of air conditioning use and the power consumption ratio between on and off states. (7)

[0034] In the formula, K is the ratio of the power consumption when the air conditioner is on to the power consumption when it is off. The probability of turning on the air conditioner at different times. The probability of turning on the air conditioner all day.

[0035] Based on the above technical concept, it should be noted that when calculating the ratio K of power consumption when the air conditioner is on to power consumption when it is off, it is necessary to rely on statistical laws and consider the ambient temperature at different times of the day. Probability function of air conditioner being turned on To improve the accuracy of the model, the relationship is determined, and the formula is as follows:

[0036] In the formula, , , , , and These are all fitting coefficients, under heating and cooling conditions. , and Different To account for the electricity consumption per 100 kilometers of air conditioning use, the unit is kWh, which includes the temperature at different times of the day. Bring into In this way, you can get the probability of turning on the air conditioner at different times. Then, calculate the probability of turning on the air conditioner all day. Therefore, the average electricity consumption per 100 kilometers on that day, taking into account the use of air conditioning, is approximately equal to the above formula (7).

[0037] Having assessed the impact of the above factors on power consumption, step S3 can then be executed: establishing a model for total power consumption per 100 kilometers to obtain the expected power consumption per 100 kilometers for the electric vehicle considering all factors. : (9).

[0038] In the formula, EC1 represents the impact of vehicle type on power consumption, EC2 represents the impact of traffic conditions on power consumption, EC3 represents the impact of aging on power consumption, and EC4 represents the impact of air conditioning use on power consumption. It is understandable that establishing a model for total power consumption per 100 kilometers considers the power consumption characteristics of different brands of electric vehicles. By statistically analyzing the proportion of electric vehicle brands in each interval and the average power consumption in each interval to calculate the expected total power consumption, this method can more accurately predict the power consumption behavior of electric vehicles and provide more accurate data support for subsequent charging load forecasting.

[0039] In one embodiment of the present invention, in step S4, the input parameters include: the number of active vehicles used to determine the total number of vehicles participating in charging within the target area. Temperature variation is used to account for the impact of ambient temperature on vehicle energy consumption and charging efficiency, using the temperature function of the day. This represents the battery capacity used to quantify the capacity of each electric vehicle, and is a factor affecting charging demand. Charging power and initial charging status Factors, due to the initial charging state The state of charge at the start of a day's activity or the end of a charge cycle, therefore =1.

[0040] In step S4, based on the above technical concept, in the process of predicting and generating the load curve of the total charging demand of all electric vehicles over time, it is first necessary to combine local electric vehicle travel statistics and the temperature of the day to determine the key parameters affecting power consumption, so as to obtain the accurate total power consumption of electric vehicles per 100 kilometers during driving. The aim is to improve forecast accuracy, that is, by accurately calculating the total power consumption per 100 kilometers, the accuracy of charging load forecasting can be improved, thereby managing the grid load more effectively; and to optimize resource allocation, that is, accurate power consumption data helps grid operators better plan and allocate resources to meet the charging needs of electric vehicles, while avoiding grid overload.

[0041] In terms of precise total power consumption After obtaining the data, during the process of predicting and generating the load curve, the following steps are performed:

[0042] S4-1. Simulate the first charging moment using the Monte Carlo method. and mileage The probability distribution is used to improve the accuracy of the model.

[0043] S4-2. Calculate the remaining battery state of charge at the time of the first charge. , (10), and based on the remaining power and charging efficiency ( Calculate the charging time of electric vehicles (specified as 0.95). , (11), where P bat For charging power, C bat Battery capacity that affects charging demand.

[0044] S4-3. By iteratively calculating and superimposing the charging load of multiple electric vehicles, a load curve is generated representing the total charging demand of all electric vehicles in the target area over time. (12), where, This represents the number of electric vehicles, and the load curve will show the charging demand at different points in time, providing data support for grid load management and the formulation of electric vehicle charging strategies.

[0045] Based on the above technical concept, it should be noted that the Monte Carlo method is used to simulate the first charging time. and mileage When the probability distribution is such that the first charging time is... and mileage The following probability distribution model needs to be satisfied because it achieves the following: a) Realistic simulation: by simulating a probability distribution that matches actual driving behavior, the charging behavior and driving habits of electric vehicle users can be more realistically reflected; b) Ensuring the activity and relevance of data, thereby improving the accuracy of the model; c) Considering randomness: the charging time and driving mileage of electric vehicles are random, and satisfying the relationship through the probability distribution model can capture this randomness, making the model closer to reality. The process is as follows:

[0046] For the first charging moment Its probability distribution must satisfy the normal distribution model: (13)

[0047] In the formula, and The starting charging time of electric vehicles The variance and standard deviation that satisfy a normal distribution;

[0048] For driving mileage Its probability distribution must satisfy the log-normal distribution model: (14)

[0049] In the formula, and Electric vehicle driving range The variance and standard deviation satisfy a normal distribution.

[0050] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A method for constructing an electric vehicle charging load prediction model considering multiple power consumption factors, characterized in that: The method comprises the steps of: By acquiring and cleaning the electric vehicle ownership and travel statistics in the target area, excluding long-term unused vehicles, and distinguishing the travel patterns during holidays and non-holidays, the number of daily active electric vehicles is determined; Analyze the multiple power consumption factors of the electric vehicle, including the brand, driving road conditions, aging degree and in-vehicle equipment usage, evaluate the influence of the factors on the power consumption, and divide the power consumption per 100 kilometers of electric vehicles of different brands into four power consumption intervals: [9kWh, 12kWh], (12kWh, 15kWh], (15kWh, 18kWh], (18kWh, 21kWh], and calculate the proportion of each interval brand, and then establish a total power consumption per 100 kilometers model according to the proportion: , wherein, is the expected value of the power consumption per 100 kilometers of the electric vehicle during driving, EC1 is the influence of the vehicle type on the power consumption, EC2 is the influence of the traffic conditions on the power consumption, EC3 is the influence of the aging degree on the power consumption, and EC4 is the influence of the air conditioner usage on the power consumption. The expected value of the electric vehicle power consumption per 100 kilometers under all factors is obtained; based on the total power consumption model, the input parameters affecting the charging load prediction are set by using the Monte Carlo simulation method, the probability distribution of the first charging time and the driving distance is simulated by using the probability distribution model, and finally the load curve representing the total charging demand of all electric vehicles in the target area over time is predicted and generated; In the process of predicting the generation of the load curve of the total charging demand of all electric vehicles over time, firstly, the key parameters affecting the power consumption need to be determined in combination with the local electric vehicle travel statistics and the temperature on the day to obtain the accurate total power consumption per 100 kilometers of the electric vehicle during driving ; secondly, after obtaining the accurate total power consumption , the following steps are continued in the process of predicting and generating the load curve. Simulating the moment of first charge using the Monte Carlo method and probability distribution of the range ​ Calculate the remaining battery state of charge at the first charging time of the day , , and according to the remaining power and charging efficiency Calculate the charging duration of the electric vehicle , Finally, by iteratively calculating the charging load of multiple electric vehicles and superimposing them, a load curve of the total charging demand of all electric vehicles in the target area over time is generated, , where, is the number of electric vehicles, P bat is the charging power, C bat is the battery capacity that affects the charging demand.

2. The method of claim 1, wherein the method further comprises: In the exclusion of long-term non-use of vehicles, define the vehicle used within two weeks as active electric vehicles, and calculate the total number of active electric vehicles based on the following formula through data obtained from surveys or relevant agencies: , wherein, is the total number of active electric vehicles; is the total number of long-term non-use electric vehicles; is the target area electric vehicle ownership; is the electric vehicle activity factor; and with 14 days as a cycle, by calculating the daily travel vehicle expectation value of holidays and non-holidays , , distinguishing holiday and non-holiday travel modes: ; , in the formula, is the number of holiday days, 14- is the number of non-holiday days.

3. The method of claim 1, wherein the method further comprises: determining a charging time of the electric vehicle; and determining a charging power of the electric vehicle. When evaluating the influence of multiple power consumption factors on power consumption, the method operates as follows: By counting the proportion of different brands of electric vehicles in each power consumption interval , , and , the power consumption per 100 kilometers is calculated to evaluate the impact of vehicle type on power consumption ; By introducing an average congestion factor , the power consumption is adjusted based on the traffic condition to evaluate the influence of the traffic condition on the power consumption EC2, , wherein, , is the average congestion distance per 100 kilometers of the electric vehicle, is the average power increase rate during congestion. By considering the aging coefficients of the battery and the motor of the electric vehicle and to quantify the influence of the aging degree on the power consumption EC3, ; By analyzing the air conditioner usage probability and its power consumption proportion in the open and closed states, the influence of air conditioner usage on power consumption is evaluated EC4, , wherein K is the ratio of power consumption when the air conditioner is open to that when the air conditioner is closed, is the probability of opening the air conditioner at different times, is the probability of opening the air conditioner all day.

4. The method of claim 1, wherein the method further comprises: The input parameters include: the number of active vehicles for determining the total number of vehicles participating in charging in the target area , temperature changes for considering the influence of ambient temperature on vehicle energy consumption and charging efficiency, using the temperature function of the day , representing the battery capacity of each electric vehicle, which affects the battery capacity of the charging demand , charging power And the initial state of charge Factor.

5. The method of claim 1, wherein the method further comprises: The first charging time is simulated using a Monte Carlo method and the probability distribution of the driving range The first charging time and the driving range need to meet the following probability distribution model to improve the accuracy of the model, wherein, For the first charging time The probability distribution thereof needs to satisfy a normal distribution model: , wherein, and are the starting charging time and the ending charging time of the electric vehicle respectively; and the variance and the standard deviation satisfying the normal distribution. For the driving range The probability distribution thereof needs to satisfy a lognormal distribution model: , wherein, and are the mean and standard deviation of the driving range of the electric vehicle satisfying the normal distribution.

6. The method of claim 3, wherein the method further comprises: When calculating the ratio K of the power consumption of the air conditioner when it is on and when it is off, the relationship between the ambient temperature at different times of the day and the air conditioner opening probability function needs to be determined according to the statistical law to improve the accuracy of the model, and the formula is as follows: , wherein a1, a2, a3, b1, b2 and b3 are fitting coefficients, b1, b2 and b3 are not the same in the heating and cooling states, is the power consumption per 100 kilometers of the air conditioner, and the unit is kWh, wherein the temperature at different times of the day is brought into , and the probability of turning on the air conditioner at different times can be obtained, and the probability of turning on the air conditioner all day is further obtained.​​​ 7. The method of claim 6, wherein the method further comprises: The ambient temperature By day weather temperature data acquisition.

Images