CN110232219B - A Data Mining-Based Approval Method for Dispatchable Capacity of Electric Vehicles - Google Patents

Abstract

The invention provides a data mining-based method for checking the dispatchable capacity of an electric vehicle. The method comprises the following steps: firstly, establishing an automobile transportation trip chain concept, and establishing a mathematical model including first trip time, travel time, parking time, travel mileage and a spatial probability transfer matrix by performing data mining on a database, and forming an electric automobile transportation trip chain; then, on the basis of the automobile traffic behavior model, a charging behavior model is established; then, simulating the traffic behavior and the charging behavior of the automobile to obtain an automobile real-time SOC curve; deducing a schedulable SOC judgment threshold and an automobile schedulable time period under different power grid frequency modulation modes; and finally, calculating the schedulable capacity of the automobile. The invention provides a data mining-based electric vehicle schedulable capacity verification method, which can provide a basis for power grid peak and frequency modulation, and schedule the charging and discharging of an electric vehicle according to the schedulable capacity in the power grid scheduling frequency modulation.

Description

A Data Mining-Based Approval Method for Dispatchable Capacity of Electric Vehicles

technical field

The invention relates to the technical field of electric vehicle intelligent control, in particular to a method for verifying the schedulable capacity of an electric vehicle.

Background technique

In recent years, due to the large demand for urban public transportation and energy and environmental issues, electric vehicles powered by green energy have become more and more popular. Electric vehicles are mainly operated in cities and can be regarded as dominant loads and hidden power sources. Their uncertain charging and discharging behaviors will have a great impact on the operation of urban power grids. Therefore, the evaluation of the charging and discharging capacity of electric vehicles is of great significance to the future urban traffic network and urban power grid.

In recent years, relevant scholars have proposed many methods to conduct a lot of research on the charging and discharging scheduling of electric vehicles. The schedulable capacity of electric vehicles should be the electric energy that meets the diverse dispatching needs of the power grid under the premise of ensuring the travel needs of users. Whether electric vehicles are dispatched is judged according to the real-time remaining power (State of Charge, SOC) of electric vehicles. Only when the SOC-related constraints are satisfied, the grid can continue to carry out capacity scheduling. Therefore, correctly and scientifically simulating the traffic behavior of electric vehicles and then obtaining their SOC curves is a prerequisite for the study of dispatchable capacity. Nowadays, most literatures usually give a certain probability distribution model when obtaining the SOC curve. This method is too simplified and cannot describe the power changes during the driving process of the car in a detailed and scientific manner, which has a significant impact on the results of subsequent schedulable capacity.

Contents of the invention

The purpose of the present invention is to solve the above-mentioned deficiencies in the schedulable capacity of electric vehicles, and provide a method for checking the schedulable capacity of electric vehicles based on data mining, so that the charging and discharging load of electric vehicles can provide reference and basis for power grid dispatching and frequency regulation.

The present invention proposes a data mining-based method for assessing the dispatchable capacity of electric vehicles, which includes the following steps:

(1) Analyze the data of car travel locations to obtain the main areas of car travel, which will be used as the basis for subsequent traffic behavior analysis;

(2) Mining and analyzing the data of the car's first trip time, driving time, parking time, driving mileage and spatial probability transfer matrix, forming a car's one-day travel chain, and deriving the car's traffic behavior model;

(3) Analyze the charging process of the car, and deduce the charging behavior model of the car.

(4) By Monte Carlo simulation method, the traffic behavior and charging behavior of the car are simulated to obtain the SOC change curve of the car for a day.

(5) According to the three frequency modulation methods of the power grid, the SOC threshold and the dispatchable period of the vehicle under the corresponding three dispatching methods are derived.

(6) According to the vehicle SOC curve and the schedulable SOC threshold, calculate the schedulable capacity of the vehicle.

In the above data mining-based verification method for the schedulable capacity of electric vehicles, the distribution ratio of the main areas of car travel is obtained by analyzing the data of car travel locations.

In the above data mining-based schedulable capacity verification method for electric vehicles, the first travel time of electric vehicles obtained by fitting the vehicle driving data satisfies the following formula (1) multidimensional normal distribution:

In the formula, where a ₁ =0.34, μ ₁ =7.46, σ ₁ =0.77; a ₂ =0.66, μ ₂ =9.20, σ ₂ =2.75

By fitting the vehicle driving data, the driving time of the electric vehicle satisfies the following formula (2) logarithmic normal distribution:

In the formula, t _tr is the travel time, μ _tr , σ _tr are the expectation and standard deviation of the corresponding starting and ending points, the values are shown in the following matrix, the i-th row represents the departure point D _i , and the j-th row and column represent the arrival point D _j .

By fitting the vehicle driving data, the parking duration of the electric vehicle satisfies the following formula (3) logarithmic normal distribution:

In the formula, t _d is the driving time, μ _td , σ _td are the expectation and standard deviation of the corresponding parking place;

Cars driving in the city can basically be regarded as driving at a constant speed, and the mileage and driving time of electric vehicles obtained by fitting the vehicle driving data can be regarded as a linear relationship.

d=v(t _tr )×t _tr (3)

According to the central limit law and the law of large numbers, it can be known that the mileage and the driving time satisfy the logarithmic normal distribution of the formula (4), and its expectation and standard deviation μ _d (t _tr ), σ _d (t _tr ) should be the same as The duration t _tr is related. Now divide t _tr into a window of 20 minutes for data analysis, and obtain the values of μ _d and σ _d in each time window.

Users can start their journey at any time and place, and their destination can also be any area of D1-D4. The choice of the destination is related to the moment when the user's journey starts and the starting place, so several spatial transition probability matrices with time as the segment can be used to describe the destination of the journey starting at the corresponding time and starting place. That is, a k*4*4 matrix P _k discretized according to a certain time period t _k is used to describe the spatial transfer situation within the time period. The i-th row of the matrix P _k represents the departure point D _i , and the j-th column represents the arrival point D _j . k is the number of time intervals after discretization, and P _k is the spatial transition probability matrix in k time periods. For example, a day is divided into k=12 time periods according to 2 hours as a time period. For example, P ₁₀ is the spatial probability transition matrix of 20-21 points.

Through the above steps, the one-day travel chain of the electric vehicle can be established, and the traffic behavior model of the vehicle can be deduced. The charging behavior of electric vehicles is now analyzed.

After the end of each journey, the car decides whether to charge according to the remaining power. Formula (5) is the criterion for car charging.

(E×SOC _i -d _i ×k)≤0.4E (5)

In the formula, E is the capacity of the car battery, kW h; d _i is the mileage of the i-th trip, km; k is the power consumption per kilometer of the car, kW h/km; SOC _i is the start of the i-th trip electricity.

After charging, the SOC _i+1 of the i+1th trip is:

In the formula, η is the charging efficiency, which is taken as 0.9. P is the slow charging or fast charging power, kW.

From the start of the car to the stop, the electric power decreases linearly according to the power consumption k per kilometer, and the SOC of the driving process is calculated as formula (7). The process from the beginning of charging to the end of charging is considered to be a constant power charging mode, and the SOC presents a linear growth trend until the end of charging. The SOC of the charging process is calculated as formula (8).

In the formula, SOC _i is the power at the beginning of the ith trip; k is the power consumption per kilometer; d _i is the mileage of the i-th trip, km; E is the battery capacity kW h; t _tr is the travel time, min.

In the formula, SOC _i0 is the power at the beginning of charging for the i-th trip; SOC _i+1 is the initial power for the i+1-th trip, and t _c is the charging time.

Through the above charging behavior analysis, the continuous and complete SOC curve of the car can be obtained.

Based on the above situation, the present invention first establishes the concept of the automobile traffic travel chain, establishes the electric vehicle travel model by data mining the database, and simultaneously establishes the electric vehicle charging model, simulates the vehicle driving process to obtain the real-time SOC curve, and then obtains the schedulable conditions The lower SOC judgment threshold, and then analyze the schedulable time period and scheduling method of the car based on this, and finally evaluate the schedulable capacity of the electric vehicle.

Compared with the prior art, the present invention has the following advantages and technical effects:

In the present invention, the schedulable capacity is calculated based on the SOC variation curve of the electric vehicle. The traffic behavior model of electric vehicles is obtained through data analysis, and a complete and continuous SOC change curve is derived by combining the charging behavior model. When establishing the traffic behavior model, the present invention considers the car travel situation in various time periods and places, and derives a relatively accurate mathematical model, which can be more suitable for the randomness of car travel in real life. According to the three frequency modulation modes of the power grid, the present invention respectively provides corresponding SOC threshold calculation methods and dispatching conditions, and can determine the maximum dispatchable potential capacity under the three situations under the condition of ensuring normal running of the automobile.

Description of drawings

Fig. 1 is the SOC change curve of the electric vehicle discharge scheduling in the embodiment.

Fig. 2 is a flow chart of the method for checking the schedulable capacity of electric vehicles in the embodiment.

Fig. 3 shows the schedulable capacity under the three scheduling modes in the embodiment.

Figure 4 shows the primary scheduling capacity of the four types of areas in the embodiment.

Fig. 5 is a schematic diagram of the overall flow of the method for verifying the dispatchable capacity of electric vehicles based on data mining.

detailed description

The specific implementation of the present invention will be further described below in conjunction with accompanying drawings and examples, but the implementation and protection of the present invention are not limited thereto. Understand or realize with reference to prior art.

Figure 5 reflects the overall flow chart of the method for checking the dispatchable capacity of electric vehicles based on data analysis. Figure 2 is a further detailed process in this embodiment, including the following steps:

(1) Analyze the data of car travel locations to obtain the main areas of car travel, which will be used as the basis for subsequent traffic behavior analysis;

(2) Mining and analyzing the data of the car's first trip time, driving time, parking time, driving mileage and spatial probability transfer matrix, forming a car's one-day travel chain, and deriving the car's traffic behavior model;

(3) Analyze the charging process of the car, and deduce the charging behavior model of the car.

(4) By Monte Carlo simulation method, the traffic behavior and charging behavior of the car are simulated to obtain the SOC change curve of the car for a day.

(5) According to the three frequency modulation methods of the power grid, the SOC threshold and the dispatchable period of the vehicle under the corresponding three dispatching methods are derived.

(6) According to the vehicle SOC curve and the schedulable SOC threshold, calculate the schedulable capacity of the vehicle.

In the above data mining-based verification method for the dispatchable capacity of electric vehicles, by analyzing the data of vehicle travel locations, the distribution ratio of the main areas of vehicle travel is obtained as follows:

Table 1 Proportion of Electric Vehicle Travel Destinations

The four types of areas are represented by D1, D2, D3, and D4 respectively.

In the above data mining-based schedulable capacity verification method for electric vehicles, the first travel time of electric vehicles obtained by fitting the vehicle driving data satisfies the following formula (1) multidimensional normal distribution:

In the formula, where a ₁ =0.34, μ ₁ =7.46, σ ₁ =0.77; a ₂ =0.66, μ ₂ =9.20, σ ₂ =2.75

By fitting the vehicle driving data, the driving time of the electric vehicle satisfies the following formula (2) logarithmic normal distribution:

In the formula, t _tr is the travel time, μ _tr , σ _tr are the expectation and standard deviation of the corresponding starting and ending points, the values are shown in the following matrix, the i-th row represents the departure point D _i , and the j-th row and column represent the arrival point D _j .

By fitting the vehicle driving data, the parking duration of the electric vehicle satisfies the following formula (3) logarithmic normal distribution:

In the formula, t _d is the driving time, μ _td , σ _td are the expectation and standard deviation of the corresponding parking place;

Table 2 Logarithmic distribution parameters of parking duration

Cars driving in the city can basically be regarded as driving at a constant speed, and the mileage and driving time of electric vehicles obtained by fitting the vehicle driving data can be regarded as a linear relationship.

d=v(t _tr )×t _tr (3)

According to the central limit law and the law of large numbers, it can be known that the mileage and the driving time satisfy the logarithmic normal distribution of the formula (4), and its expectation and standard deviation μ _d (t _tr ), σ _d (t _tr ) should be the same as The duration t _tr is related. Now divide t _tr into a window of 20 minutes for data analysis, and obtain the values of μ _d and σ _d in each time window. The specific parameter values are shown in Table 3.

Table 3 Logarithmic distribution parameters of mileage

Users can start their journey at any time and place, and their destination can also be any area of D1-D4. The choice of the destination is related to the moment when the user's journey starts and the starting place, so several spatial transition probability matrices with time as the segment can be used to describe the destination of the journey starting at the corresponding time and starting place. That is, a k*4*4 matrix P _k discretized according to a certain time period t _k is used to describe the spatial transfer situation within the time period. The i-th row of the matrix P _k represents the departure point D _i , and the j-th column represents the arrival point D _j . k is the number of time intervals after discretization, and P _k is the spatial transition probability matrix in k time periods. For example, a day is divided into k=12 time periods according to 2 hours as a time period. For example, P ₁₀ is the spatial probability transition matrix of 20-21 points.

Through the above steps, the one-day travel chain of the electric vehicle can be established, and the traffic behavior model of the vehicle can be deduced. The charging behavior of electric vehicles is now analyzed.

All spatial probability transition matrices are as follows.

After the end of each journey, the car decides whether to charge according to the remaining power. Formula (5) is the criterion for car charging.

(E×SOC _i -d _i ×k)≤0.4E (5)

In the formula, E is the capacity of the car battery, kW h; d _i is the mileage of the i-th trip, km; k is the power consumption per kilometer of the car, kW h/km; SOC _i is the start of the i-th trip electricity.

After charging, the SOC _i+1 of the i+1th trip is:

In the formula, η is the charging efficiency, which is taken as 0.9. P is the slow charging or fast charging power, kW.

From the start of the car to the stop, the electric power decreases linearly according to the power consumption k per kilometer, and the SOC of the driving process is calculated as formula (7). The process from the beginning of charging to the end of charging is considered to be a constant power charging mode, and the SOC presents a linear growth trend until the end of charging. The SOC of the charging process is calculated as formula (8).

In the formula, SOC _i is the power at the beginning of the ith trip; k is the power consumption per kilometer; d _i is the mileage of the i-th trip, km; E is the battery capacity kW h; t _tr is the travel time, min.

In the formula, SOC _i0 is the power at the beginning of charging for the i-th trip; SOC _i+1 is the initial power for the i+1-th trip, and t _c is the charging time.

Through the above charging behavior analysis, the continuous and complete SOC curve of the car can be obtained.

In the above-mentioned data mining-based verification method for the dispatchable capacity of electric vehicles, when the electric vehicles are in the idle state connected to the network, that is, when the SOC curve remains horizontal, the electric vehicles have the basic conditions for capacity dispatching. At the same time, the charging and discharging scheduling of the car must also be based on the condition that the next trip will not be affected, so more detailed SOC discrimination constraints should be set for this interval.

(1) Discharge scheduling SOC threshold calculation

Figure 1 shows the SOC variation curve of electric vehicle discharge dispatching. When the car is in an idle state and the SOC value is high, the car has the possibility of discharge scheduling at this time. However, the discharge scheduling should ensure that the remaining power of the car after the discharge is completed will not affect the next trip. For example, the SOC value at the beginning of each trip should not be less than 0.2, and the calculation formula of the discharge scheduling SOC threshold is (9-11).

In the formula, SOC ⁰ _i+1 is the power at the end of the i-th trip scheduling; SOC _iend is the power after the i-th trip car is connected to the grid and responds to the discharge dispatch; SOC _lim is 0.2; P is the charging power of the grid, kW; t _c is the charging time from the end of the car's discharge to the time when it is off-grid, min; E is the battery capacity kW·h.

In the formula, SOC _i ⁰ is the electric quantity at the beginning of the i-th trip scheduling; p is the discharge power of the vehicle, kW; t _d is the scheduling time, min.

The grid-connected power of the car should be greater than the discharge threshold SOC _{dis_i} , and the constraints of discharge scheduling are:

(2) Charging scheduling SOC threshold calculation

When the electric vehicle is in the idle state of the network, as long as the battery is not fully charged, it has the potential of charging scheduling. Therefore, the charging scheduling SOC threshold is obtained by the following formula (12):

In the formula, SOC ⁰ _i+1 is the power at the end of the i-th trip scheduling; t _d is the scheduling time, min.

The grid-connected power of the car should be less than the charging threshold SOC _{charge_i} , and the constraints for charging scheduling are:

In the above data mining-based verification method for the dispatchable capacity of electric vehicles, affected by the three frequency modulation methods of the power grid, the dispatching time of different dispatching methods of electric vehicles is different, and the dispatching time is shown in Table 3:

Table 3 Scheduling time

If the idle time of the car connected to the network is not shorter than the schedulable time, then the value of t _d in the formula is taken according to the above table, otherwise, the schedulable condition is not met.

In the above data mining-based verification method for dispatchable capacity of electric vehicles, the calculation of dispatchable capacity of electric vehicles meets the following requirements.

To simplify the analysis, the charging and discharging process of electric vehicles is regarded as constant power charging and discharging (p).

When the electric vehicle is in the grid-connected idle state, that is, when the SOC curve remains horizontal, it is satisfied: SOC _{dis_i} ≤ SOC _i (t) ≤ 1;

Then the discharge dispatch capacity P _discharge (t)=p.

When the electric vehicle is in the grid-connected idle state, that is, when the SOC curve remains horizontal, it is satisfied: SOC _i (t) ≤ SOC _{charge_i} ≤ 1;

Then the charging scheduling capacity P _charge (t)=p.

When the electric vehicle SOC does not meet the above conditions, or the idle time of the network connection is longer than the dispatchable time, the charging and discharging dispatching potential is 0.

P _dis (t) = 0; P _charge (t) = 0.

The sum of the dispatchable capacity of all electric vehicles in the grid satisfies formula (14):

Fig. 3 shows the schedulable capacity of three dispatching modes for 300,000 electric vehicles in this embodiment, and scenarios 1, 2 and 3 represent the primary, secondary and tertiary dispatching modes respectively. Figure 4 shows the one-time dispatch capacity of 300,000 electric vehicles in four types of areas, which can be dispatched according to the dispatchable capacity in power grid dispatch and frequency regulation.

It can be seen from Figure 3 that the dispatchable potential capacity of the car decreases with the increase of the dispatching time. When the dispatching time is short in the first dispatching and the second dispatching, the power consumed by the car’s previous driving has a strong dispatching effect when the car is parked for a short time. potential. The time for three dispatches is as long as 120 minutes. At this time, the idle time of network access and parking that meet the three dispatches are more than 120 minutes. This kind of long-term parking is generally for resting and parking at night. The potential is very limited, so the curves in (a) have very small values.

The longer the dispatch time, the car has enough time to recharge after discharge, so the calculated discharge SOC threshold is lower, and a large number of cars in the grid have the potential for discharge dispatch, so in (b), the curve value of the third case remains at a high level. Generally speaking, in any dispatching situation, the dispatchable potential capacity of discharge in the power grid is relatively high, and the response to the change of dispatching time is not obvious, while the dispatchable potential capacity of charging is high in the short dispatching time, and has a large impact on dispatching time. The change response is obvious.

The schedulable potential capacity of the first and second charging gradually increases to the maximum at night, because after a day of driving, the power of the car at night is at a low level. The schedulable potential capacity of discharge gradually increases from early morning to morning, because the car is fully charged and starts to travel during this period, and the power is at a relatively high level. In general, the dispatchable potential capacity of discharge throughout the day can basically cover the dispatchable potential capacity of charging, and there is still a residual value to supplement other loads of the power grid, which will bring great benefits to the operation of the power grid.

According to the previous simulation of the travel behavior of electric vehicles, the car has four types of areas as destinations according to the relevant travel characteristics, and the charging behavior of the car in different areas is different. Therefore, the dispatchable potential capacity in different regions is also different. The primary dispatchable potential capacity curves of the four types of regions are obtained, which can be used to analyze the regional differences in dispatchable potential capacity.

It can be seen from Figure 4 that the charge and discharge potential capacity of electric vehicles in the four types of areas has a similar change with time, and the charge and discharge potential capacity is concentrated in the relevant time period when the vehicle is in the area, and the capacity time distribution in different areas is relatively large. the difference. The dispatching potential of the living quarters is large mainly during the late night to early morning hours, and the discharge capacity of the car can be used for the remaining loads of the grid as electric energy. The working area has a great discharge potential during the day, which can provide extra power for the grid for the rest of the loads. Centralized dispatching of the discharge load in the working area can make full use of power resources and improve efficiency. The charging and discharging potential of commercial and leisure areas is concentrated in the afternoon and evening. Scheduling the discharge capacity during this period can compensate for the large load demand of commercial and leisure areas nearby. To sum up, targeted power grid dispatching in different regions and time periods can greatly tap the hidden power potential of shared cars, while improving the stability and flexibility of power grid operation.

The above-mentioned embodiment is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited by the above-mentioned embodiment, and any other modifications, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principles of the present invention , should be equivalent replacement methods, and should be included within the protection scope of the present invention.

Claims (6)

1. A method for checking and approving the dispatchable capacity of electric vehicles based on data mining is characterized in that it comprises the following steps: (1) Analyze the data of car travel locations to obtain the main areas of car travel; (2) Mining and analyzing the data of the car's first trip time, driving time, parking time, driving mileage and spatial probability transfer matrix to form a car's one-day travel chain and establish a car traffic behavior model; (3) Analyze the charging process of the car and establish the charging behavior model of the car, including: After the end of each journey, the car decides whether to charge according to the remaining power. The formula (5) is the criterion for charging the car. (E×SOC _i -d _i ×k)≤0.4E In the formula, E is the capacity of the car battery, kW h; d _i is the mileage of the i-th trip, km; k is the power consumption per kilometer of the car, kW h/km; SOC _i is the start of the i-th trip electricity; After charging, the SOC _i+1 of the i+1th trip is:

In the formula, η is the charging efficiency, which is taken as 0.9, P is the slow charging or fast charging power, and the unit is kW; t _c is the charging time; When the car starts to stop, the electric power decreases linearly according to the power consumption k per kilometer. The SOC calculation of the driving process is as follows:

In the formula, SOC(t) is the power at any time t, SOC _i is the power at the beginning of the i-th trip; k is the power consumption per kilometer; d _i is the i-th driving mileage, unit km; E is the battery capacity kW·h; t _tr is the driving time, unit min; From the beginning of charging to the end of charging, the car is considered to adopt the constant power charging mode, and the SOC presents a linear growth trend until the end of charging. The specific value of SOC during the charging process is calculated as follows:

In the formula, SOC _i0 is the power when the i-th trip starts to charge; SOC _i+1 is the initial power of the i+1-th trip; (4) Through the Monte Carlo simulation method, the traffic behavior and charging behavior of the car are simulated to obtain the SOC change curve of the car for a day; (5) According to the three frequency modulation methods of the power grid, the SOC threshold and the dispatchable period of the vehicle under the corresponding three dispatching methods are derived; (6) According to the vehicle SOC curve and the schedulable SOC threshold, the schedulable capacity of the vehicle is derived, and the charging and discharging capacity of the electric vehicle is verified under different frequency modulation methods of the power grid. 2. The method for verifying the dispatchable capacity of electric vehicles based on data mining according to claim 1, characterized in that: When choosing a car travel destination, according to the time period of the car travel time and the difference in the starting location, use a k*4*4 matrix P _k discretized according to the set time period t _k to describe the spatial transfer in this time period Situation; the i-th row of the matrix P _k represents the departure point D _i , and the j-th column represents the arrival point D _j ; k is the number of time intervals after discretization, and P _k is the spatial transition probability matrix within the k period. 3. the electric vehicle schedulable capacity verification method based on data mining according to claim 2, is characterized in that: take tk=2 hours as the interval to form the spatial probability transfer matrix of _k =12 time periods in one day; according to this matrix, The car selects the destination to depart at the corresponding time and place; If P ₁₀ is the spatial probability transition matrix of 20-21 points;

4. the electric vehicle schedulable capacity verification method based on data mining according to claim 2, is characterized in that the traffic behavior model of the automobile set up in step (2) is as follows: The first travel time of the car satisfies the following multidimensional normal distribution:

In the formula, where a ₁ =0.34, μ ₁ =7.46, σ ₁ =0.77; a ₂ =0.66, μ ₂ =9.20, σ ₂ =2.75; The driving time of the car satisfies the following lognormal distribution:

In the formula, t _tr is the driving time, t _d is the parking time, μ _tr and σ _tr are the expectation and standard deviation of the corresponding starting and ending points, the values are shown in the following matrix, the i row represents the departure point D _i , and the j row and column represent the arrival ground D _j ;

The car parking time satisfies the lognormal distribution of the following formula:

In the formula, μ _td and σ _td are the expectation and standard deviation of the corresponding parking place; The mileage of the car satisfies the logarithmic normal distribution within a certain period of time window, and 20 minutes is divided into a window for data analysis:

In the formula, d is the mileage, μ _d , σ _d are the expectation and standard deviation of the mileage in each time window. 5. The data mining-based electric vehicle schedulable capacity verification method according to claim 1, characterized in that step (5) comprises: When the SOC curve of the car is horizontal, the car is in the idle state of the grid; according to the three frequency modulation methods of the power grid, three scheduling methods are set, and the scheduling time is 15 minutes, 30 minutes, and 120 minutes; The calculation method of the discharge scheduling SOC threshold criterion is as follows:

In the formula, SOC ⁰ _i+1 is the power at the end of the i-th trip scheduling; SOC _iend is the power after the i-th trip car is connected to the grid and responds to the discharge dispatch; SOC _lim is 0.2; P is the charging power of the grid, in kW; t _c is the charging time from the end of the car's discharge to the time when it is off-grid, in min; E is the battery capacity kW h;

In the formula, SOC _i ⁰ is the electric quantity at the beginning of the scheduling of the i-th trip; p is the discharge power of the vehicle, kW; t _d is the scheduling time, the unit is min; The grid-connected power of the car should be greater than the discharge threshold SOC _{dis_i} , and the constraints of discharge scheduling are:

The determination of the charging scheduling SOC threshold criterion is as follows: When the electric vehicle is in the idle state of the grid, as long as the battery is not fully charged, it has the potential for charging scheduling. Therefore, the charging scheduling SOC threshold is obtained by the following formula:

In the formula, SOC ⁰ _i+1 is the power at the end of the i-th trip scheduling; t _d is the scheduling time, min; SOC _{charge_i} indicates the charging threshold; The grid-connected power of the car should be less than the charging threshold SOC _{charge_i} , and the constraints for charging scheduling are:

6. The method for verifying the schedulable capacity of electric vehicles based on data mining according to claim 1, wherein step (6) specifically comprises: The charging and discharging process of electric vehicles is regarded as constant power p charging and discharging, and the schedulable evaluation process of electric vehicles is: When the electric vehicle is in the grid-connected idle state, that is, when the SOC curve remains horizontal, it is satisfied: SOC _{dis_i} ≤ SOC _i (t) ≤ 1; Then the discharge dispatching capacity P _discharge (t)=p; When the electric vehicle is in the grid-connected idle state, that is, when the SOC curve remains horizontal, it is satisfied: SOC _i (t) ≤ SOC _{charge_i} ≤ 1, and SOC _{charge_i} represents the charging threshold; Then the charging scheduling capacity P _charge (t)=p; When the electric vehicle SOC does not meet the above conditions, or the idle time of the network connection is longer than the dispatchable time, the charging and discharging dispatching potential is 0; P _dis (t) = 0; P _charge (t) = 0; For the power grid, the schedulable capacity of a single electric vehicle is insignificant, and the schedulable charge and discharge capacity should be equal to the sum of the dispatchable capacity of electric vehicles in the distribution network:

Images