Electric vehicle group aggregation load prediction method, medium and equipment

Abstract

The invention provides an electric vehicle group aggregation load prediction method, a medium and equipment, and relates to the technical field of electric vehicle network interconnection, and the method comprises the steps: predicting the number sequence of electric vehicles arriving at a charging station in a time sequence through Poisson distribution, constructing an electric vehicle charging state double-layer discretization model based on a high-order Markov chain model, and calculating a transition probability matrix estimation value and a coefficient of the high-order Markov chain model according to the electric vehicle number sequence and the electric vehicle charging state double-layer discretization model, and further determining a charging load prediction value. According to the method, the aggregation load of the electric vehicle group can be accurately predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle network interconnection, and particularly to a method, medium and device for predicting the aggregated load of an electric vehicle group. Background Art

[0002] Electric vehicles with mobile energy storage characteristics are a type of flexible and high-quality demand-side resource. Due to the reduction of battery costs for electric vehicles, environmental benefits, and electricity prices that are cheaper compared to the rising prices of fossil fuels, the electrification of transportation has become an increasing trend. However, the large-scale popularization of the behavior of electric vehicles without effective charging management may have a negative impact on the existing power transmission and distribution networks, such as forming peaks, increasing power fluctuations, poor power quality, and voltage over-limit. Reasonable use of controlling the charging time and charging times of electric vehicles and other characteristics can reduce their impact on the power system. Therefore, accurately predicting the aggregated load model of an electric vehicle cluster is an important means to realize electric vehicles during peak and trough periods of the grid load and other regulation demands. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, medium and device for predicting the aggregated load of an electric vehicle group, aiming to accurately predict the aggregated load of an electric vehicle group. The specific technical solutions are as follows:

[0004] A method for predicting the aggregated load of an electric vehicle group, the method comprising the following steps:

[0005] S100. Predict the sequence of the number of electric vehicles arriving at the charging station in the time series through Poisson distribution;

[0006] S200. Construct a two-layer discretization model of the charging state of electric vehicles based on a high-order Markov chain model;

[0007] S300. Calculate the estimated value of the transition probability matrix and the coefficients of the high-order Markov chain model according to the sequence of the number of electric vehicles and the two-layer discretization model of the charging state of electric vehicles, and then determine the predicted value of the charging load.

[0008] Further, S100 includes the following steps:

[0009] S110. Let n t be the total number of electric vehicles that have entered the charging station for charging services as of time t, which follows a Poisson distribution with parameters:

[0010]

[0011] where P(n t = m) represents the probability, and λ represents the number of electric vehicles arriving at the charging station per unit time;

[0012] S120. Generate a sequence of electric vehicles arriving at the charging station for charging services within the nth unit time period according to the properties of the Poisson process. The expression is as follows:

[0013]

[0014] where K(t) is a time series randomly generated for the number of electric vehicles arriving within the total time, k n is the number of electric vehicles arriving at the charging station in the nth time slot, n is the number of segments per unit time within the total time period, T2 interval is the length of the unit time period, and T1 is the length of the total time period;

[0015] S130. Assume that the vehicle can enter the charging station without queuing and charging service can be provided immediately. Then the time when the electric vehicle arrives at the charging station at this time is the charging start time. The calculation formula for the charging start time is as follows:

[0016]

[0017] where T3 is the charging start time, n' is the serial number of the time period, and m' is the sequence value of the electric vehicle arriving at the nth time slot;

[0018] S140. Calculate the expected value X of the Poisson distribution. The formula is as follows:

[0019]

[0020] Furthermore, S200 includes the following steps:

[0021] S210. High-order Markov description of the charging process;

[0022] S220. Discretization of the charging state.

[0023] Furthermore, S210 includes the following steps:

[0024] S2101. The discrete-time recursive formula for the remaining battery power is expressed as:

[0025]

[0026] where S(t + 1) and S(t - h + 1) represent the state of charge at times t + 1 and t - h + 1 respectively, h represents the order, Pc represents the charging power, η represents the charging efficiency, C B is the actual capacity of the battery, and Δt is the time interval;

[0027] S2102. Discretize the state of charge of the battery and divide it into multiple state spaces, and represent the load process chain of each state interval based on the high-order Markov chain model for dynamic conversion.

[0028] Further, S220 includes the following steps:

[0029] S2201. Consider the state of charge of the battery as a fuzzy concept. The percentage of the state of charge ranges from [0, 100], and its three - quantiles are divided into three large intervals U1, U2, and U3. U1 is the small - battery - capacity state [0, 1], U2 is the normal - battery - capacity state [1, 80], and U3 is the large - battery - capacity state [80, 100].

[0030] S2202. The second - level discretization continues on the basis of the three large intervals U1, U2, and U3. Each large interval is further divided into n small intervals, and S′(i + 1) represents the lower limit value of the state of charge of the (i + 1) - th large interval, and S′(i, j) represents the lower limit value of the state of charge of the j - th small interval in the i - th large interval.

[0031] Further, S300 includes the following steps:

[0032] S310. Let the matrix composed of the second - order transition probabilities from the state interval i to the state interval j of the state - of - charge sequence be:

[0033]

[0034] Where, represents the probability value from the small - battery - capacity state to the small - battery - capacity state; represents the probability value from the small - battery - capacity state to the normal - battery - capacity state; represents the probability value from the small - battery - capacity state to the large - battery - capacity state; represents the probability value from the normal - battery - capacity state to the small - battery - capacity state; represents the probability value of the normal - battery - capacity state to the normal - battery - capacity state; represents the probability value from the normal - battery - capacity state to the large - battery - capacity state; represents the probability value from the large - battery - capacity state to the small - battery - capacity state; represents the probability value from the large - battery - capacity state to the normal - battery - capacity state; represents the probability value of the large - battery - capacity state to the large - battery - capacity state;

[0035] S320. The estimated value of F h is defined as The calculation formula is as follows:

[0036]

[0037] Where, represents the estimated value from the small - battery - capacity state to the small - battery - capacity state, represents the estimated value from the small - battery - capacity state to the normal - battery - capacity state; Represents the estimated value from the small battery capacity state to the large battery capacity state; Represents the estimated value from the normal battery capacity state to the small battery capacity state; Represents the estimated value from the normal battery capacity state to the normal battery capacity state; Represents the estimated value from the normal battery capacity state to the large battery capacity state; Represents the estimated value from the large battery capacity to the small battery capacity state; Represents the estimated value from the large battery capacity state to the normal battery capacity state; Represents the estimated value from the large battery capacity state to the large battery capacity state;

[0038] S330. Let the charging state sequence converge to the Poisson distribution. The occurrence ratio of each state in the expected value X sequence is calculated as the estimated value of the Poisson distribution expected value X Then, the coefficient λ of the high-order Markov chain model h The estimated value is calculated by the following formula:

[0039]

[0040] where, min||·|| ∞ Represents the infinity norm of the vector, and λ h Satisfies two conditions: ∑λ h = 1 and λ h ≥ 0;

[0041] S340. Obtain the electric vehicle charging load value prediction model as follows:

[0042]

[0043] where, X c (t) = [X c (t, 1), X c (t, 2), X c (t, 3)] T Is a 3-dimensional column vector representing the load value at time t, X c (t + 1) represents the predicted value of the electric vehicle load at the charging station at time t + 1, L is the maximum value of the order of the Markov chain, here L = 2 is taken, X c (t - h + 1) represents the electric vehicle charging load in the charging station at time t - h + 1, Y c (t) represents the output power of the electric vehicle at time t, C c Is the output matrix, which is a 3-dimensional unit row vector, V c (t) is the load change amount caused by external factors, with the same structure as X c (t), P' maxis the maximum charging power;

[0044]

[0045] Among them, and are the number of electric vehicles inserted and removed within the t-th time interval, and N ev (t) is the total number of electric vehicles in the charging station at time t, and X in (t) is the charging power within the t-th time interval, and X out (t) is the discharging power within the t-th time interval.

[0046] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the electric vehicle group aggregation load prediction method described above are implemented.

[0047] The present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the electric vehicle group aggregation load prediction method described above are implemented.

[0048] An electric vehicle group aggregation load prediction method, medium, and device provided by the present invention have the following

[0049] Beneficial effects:

[0050] The electric vehicle group aggregation load prediction method provided by the present invention predicts the sequence of the number of electric vehicles arriving at the charging station in the time series through Poisson distribution, constructs a two-layer discretization model of the charging state of electric vehicles based on a high-order Markov chain model, calculates the estimated value of the transition probability matrix and the coefficients of the high-order Markov chain model according to the sequence of the number of electric vehicles and the two-layer discretization model of the charging state of electric vehicles, and further determines the predicted value of the charging load amount, and can accurately predict the load aggregated by the electric vehicle group. Description of the Drawings

[0051] Figure 1 is a schematic flow chart of an electric vehicle group aggregation load prediction method provided by an embodiment of the present invention;

[0052] Figure 2 is the probability density graph of the starting time of electric vehicle charging in the verification example of the present invention;

[0053] Figure 3 is the scatter plot distribution of the electric vehicle charging start time series in the verification example of the present invention;

[0054] Figure 4 is the predicted charging station load curve under two methods in the verification example of the present invention;

[0055] Figure 5 It is the base load diagram of the distribution network in the verification example of the present invention;

[0056] Figure 6 It is the structural block diagram of the computer device according to the embodiment of the present invention. Specific embodiments

[0057] Next, in combination with the drawings provided by the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the drawings are all in a very simplified form and use non-precise scales, only for the purpose of facilitating and clearly assisting in explaining the purpose of the embodiments of the present invention.

[0058] Embodiment 1

[0059] This embodiment provides an aggregated load prediction method for an electric vehicle group. Refer to Figure 1 As shown, the method includes the following steps:

[0060] S100. Predict the sequence of the number of electric vehicles arriving at the charging station in the time series through the Poisson distribution.

[0061] The behavior of electric vehicles arriving at the charging station in sequence is described by the Poisson process: in the plug-in charging mode, the vehicle arrival time is the start time of electric vehicle charging. The Poisson distribution is an independent increment process. The charging station and the incoming electric vehicles constitute a random service system, which satisfies the following three conditions:

[0062] (1) Stability: The probability of k customers arriving within a certain time interval is independent of the start time of the interval, and only related to the length t of the interval and the number k of arriving customers. The probability is P k (t).

[0063] (2) Lack of posteriority: The number of customer arrivals within non-overlapping time intervals is independent of each other.

[0064] (3) Finiteness: The probability of a finite number of customers arriving within a finite interval of any time length is 1. The behavior of electric vehicles entering the charging station for charging obviously satisfies the above three conditions. Therefore, it can be described by the Poisson distribution. Regarding the arrival of each electric vehicle as an event, the arrival time of the i-th electric vehicle at the charging station is represented as τ i , where i = 1, 2, 3, 4...

[0065] Specifically, S100 includes the following steps:

[0066] S110. Let n t$N(t)$ is the total number of electric vehicles entering the charging station for charging services by time $t$, which follows a Poisson distribution with parameter:

[0067]

[0068] where $P(n$ t $=m)$ represents the probability, and $\lambda$ represents the number of electric vehicles arriving at the charging station per unit time;

[0069] S120. Generate the sequence of electric vehicles arriving at the charging station for charging services in the $n$-th unit time period according to the properties of the Poisson process. The expression is as follows:

[0070]

[0071] where $K(t)$ is the time series randomly generated by the total number of electric vehicles arriving within the total time, $k$ n is the number of electric vehicles arriving at the charging station in the $n$-th time slot, $n$ is the number of segments per unit time within the total time period, $T_2$ interval is the length of the unit time period, and $T_1$ is the length of the total time period;

[0072] S130. Assume that the vehicle can enter the charging station without queuing and charging service is provided immediately. Then the arrival time of the electric vehicle at the charging station at this time is the charging start time. The calculation formula for the charging start time is as follows:

[0073]

[0074] where $T_3$ is the charging start time, $n'$ is the serial number of the time period, and $m'$ is the sequence value of the electric vehicle arriving at the $n$-th time slot;

[0075] S140. Calculate the expected value $X$ of the Poisson distribution. The formula is as follows:

[0076]

[0077] S200. Construct a two-layer discretization model for the charging state of electric vehicles based on the high-order Markov chain model.

[0078] When the conditional probability distribution of the future state depends only on the current state given the current state and all past states, the stochastic process has the Markov property. When the conditional probability distribution of the future state depends only on the current state given the current state and all past states, the stochastic process has the Markov property, and it is conditionally independent of the past states. Let $\{Y(t), t\geq1\}$ be a stochastic process, then it can be represented by formula (5):

[0079] $P(Y$ t $|Y_0,Y_1,\cdots,Y$ t-1 ) = $P(Y$ t|Y t-1 ), t = 1, 2,... (5)

[0080]

[0081] Equation (6) is an H-order Markov chain model. When the parameter λ h ≧0 is the coefficient of the high-order Markov chain model, P h is the h-step transition probability matrix.

[0082] S200 includes the following steps:

[0083] S210, High-order Markov description of the charging process, including:

[0084] S2101, The discrete-time recurrence formula for the remaining battery charge is expressed as:

[0085]

[0086] where S(t + 1) and S(t - h + 1) represent the state of charge at times t + 1 and t - h + 1 respectively, h represents the order, Pc represents the charging power, η represents the charging efficiency, C B is the actual capacity of the battery, and Δt is the time interval.

[0087] The charging of an electric vehicle is a dynamic change of the battery from a low state of charge to a high state of charge and a process of dynamic change from a high state of charge to a low state of charge. The probability distribution of S(t + 1) is independent of the historical state of the battery and only depends on the state of the battery at time t - h + 1, which conforms to the Markov chain.

[0088] S2102, The state of charge of the battery is discretized into multiple state spaces and represents the load process chain of each state interval based on the high-order Markov chain model. P i,j represents the transition probability from interval i to interval j and can be expressed as the following formula using conditional probability:

[0089]

[0090] According to Equation (7), the state of charge of the battery can be regarded as a discrete random process. Thus, the state of charge of the battery is discretized into multiple state spaces and represents the load process chain of each state area based on the high-order Markov model.

[0091] S220, Discretization of the charging state, including:

[0092] S2201. Regard the state of charge of the battery as a fuzzy concept. The percentage of the state of charge ranges from [0, 100], and its three - quartiles are divided into three large intervals U1, U2, and U3. U1 is the small - battery - capacity state [0, 1], U2 is the normal - battery - capacity state [1, 80], and U3 is the large - battery - capacity state [80, 100].

[0093] Define the fuzzy set: A fuzzy set is the entire set composed of objects with specific fuzzy attributes. The definition of a fuzzy set: Given a certain non - empty domain U, let σ a be any mapping from U to [0, 1].

[0094] σ A : U → [0, 1], y → σ A (y) (9)

[0095] where A is a certain fuzzy subset of U, and σ A is the membership function of A, and σ A (y) is the membership degree of y to A. The closer σ A (y) is to 1, the higher the degree that y belongs to A. The closer σ A (y) is to 0, the lower the degree that y belongs to A.

[0096] Regard the state of charge of the battery as a fuzzy concept. The percentage of the state of charge ranges from [0, 100], and its three - quartiles are divided into three elements U1 - U3, which are the small - battery - capacity state [0, 1], the normal - battery - capacity state [1, 80], and the full - state [80, 100] respectively. They correspond to three groups of fuzzy small - battery - capacity A1, normal - battery - capacity A2, and large - battery - capacity A3. The batteries in A1 - A3 can be regarded as balanced loads and can delay charging during peak hours.

[0097] S2202. The second - level discretization continues on the basis of the three large intervals U1, U2, and U3. Each large interval is subdivided into n small intervals, and S′(i + 1) is used to represent the lower limit value of the state of charge of the (i + 1) - th large interval, and S′(i, j) is used to represent the lower limit value of the state of charge of the j - th small interval in the i - th large interval.

[0098] After the state of charge undergoes the above - mentioned discretization process, the charging process of the electric vehicle is described as a Markov process.

[0099] S300. Calculate the estimated value of the transition probability matrix and the coefficients of the high - order Markov chain model according to the double - layer discretization model of the electric - vehicle quantity sequence and the electric - vehicle charging state, and then determine the predicted value of the charging load.

[0100] Specifically, S300 includes the following steps:

[0101] S310. Let the matrix composed of the second-order transition probabilities from state interval i to state interval j in the charging state sequence be:

[0102]

[0103] Where, represents the probability value from the small battery capacity state to the small battery capacity state; represents the probability value from the small battery capacity state to the normal battery capacity state; represents the probability value from the small battery capacity state to the large battery capacity state; represents the probability value from the normal battery capacity state to the small battery capacity state; represents the probability value of the normal battery capacity state to the normal battery capacity state; represents the probability value from the normal battery capacity state to the large battery capacity state; represents the probability value from the large battery capacity to the small battery capacity state; represents the probability value from the large battery capacity state to the normal battery capacity state; represents the probability value of the large battery capacity state to the large battery capacity state;

[0104] S320. The estimated value of F h is defined as The calculation formula is as follows:

[0105]

[0106] Where, represents the estimated value from the small battery capacity state to the small battery capacity state, represents the estimated value from the small battery capacity state to the normal battery capacity state; represents the estimated value from the small battery capacity state to the large battery capacity state; represents the estimated value from the normal battery capacity state to the small battery capacity state; represents the estimated value of the normal battery capacity state to the normal battery capacity state; represents the estimated value from the normal battery capacity state to the large battery capacity state; represents the estimated value from the large battery capacity to the small battery capacity state; represents the estimated value from the large battery capacity state to the normal battery capacity state; represents the estimated value of the large battery capacity state to the large battery capacity state;

[0107] S330. Let the proportion of the occurrence of each state in the charging state sequence converging to the Poisson distribution expected value X sequence be calculated as the estimated value of the Poisson distribution expected value X Then, the coefficient λ of the high-order Markov chain model hThe estimated value is calculated by the following formula:

[0108]

[0109] where min||·|| ∞ represents the infinity norm of a vector, and λ h satisfies two conditions: ∑λ h = 1 and λ h ≥ 0;

[0110] Combined with formula (6), the prediction model of the electric vehicle charging load value is obtained as follows:

[0111]

[0112] where X c (t) = [X c (t, 1), X c (t, 2), X c (t, 3)] T is a 3-dimensional column vector representing the load value at time t, X c (t + 1) represents the predicted value of the electric vehicle load at the charging station at time t + 1, L is the maximum value of the order of the Markov chain, here L = 2 is taken, X c (t - h + 1) represents the electric vehicle charging load in the charging station at time t - h + 1, Y c (t) represents the output power of the electric vehicle at time t, C c is the output matrix, which is a 3-dimensional unit row vector, V c (t) is the load change amount caused by external factors, having the same structure as X c (t), and P' max is the maximum charging power;

[0113]

[0114] where, and are the number of electric vehicles inserted and removed within the t-th time interval, N ev (t) is the total number of electric vehicles in the charging station at time t, X in (t) is the charging power in the t-th time interval, and X out (t) is the discharging power in the t-th time interval.

[0115] The electric vehicle group aggregated load prediction method provided by the present invention predicts the sequence of the number of electric vehicles arriving at the charging station in the time series through Poisson distribution, constructs a double-layer discretization model of the charging state of electric vehicles based on a high-order Markov chain model, calculates the estimated value of the transition probability matrix and the coefficients of the high-order Markov chain model according to the sequence of the number of electric vehicles and the double-layer discretization model of the charging state of electric vehicles, and further determines the predicted value of the charging load, and can accurately predict the load aggregated by the electric vehicle group.

[0116] Verification example

[0117] Based on the data of a charging station in a logistics park, the constructed load model was verified in Matlab. This charging station has a total of 200 charging piles and can supply power to 1000 pure electric heavy trucks. Taking the 24-hour operation of the charging station as an example, it is assumed that the charging behavior of pure electric heavy trucks only occurs when the power is less than 20%, and the heterogeneity of the battery is not considered. The parameters of pure electric heavy trucks are shown in Table 1.

[0118] Table 1 Parameters of pure electric heavy trucks

[0119] Parameter Name Parameter Value Battery Capacity / kWh 60 Planned Charge and Discharge Power / kW 7 Charging Efficiency 0.94

[0120] According to the actual situation of the charging station, a total of 1000 vehicles visit the charging station for charging in a day. According to the maximum likelihood estimation, its arrival λ = 5.6 is a Poisson process. The arrival probability distribution of pure electric heavy trucks is as Figure 2 shown, and the scatter distribution of the start time series of electric vehicle charging is as Figure 3 shown, mainly concentrated in the time period from 18:00 to 21:00 when the charging starts, and the number of pure electric heavy trucks that need to be charged is small.

[0121] According to the charging start data, the predicted load curve of the charging station is calculated by formula (14) and compared with the power of Monte Carlo simulation, as Figure 4 shown. It can be seen that the charging power of the charging station fluctuates from 0 to 500 kw in the time period from 00:00 to 10:00, the charging power gradually increases in the time period from 10:00 to 20:00, reaches the peak at about 7:00, and the power gradually decreases after 20:00. It can be seen that the charging power of the charging station fluctuates from 0 to 500 kw in the time period from 00:00 to 10:00, the charging power gradually increases in the time period from 10:00 to 20:00, reaches the peak at about 7:00, and the power gradually decreases after 20:00. The power obtained by the charging load aggregation model based on the high-order Markov chain and the Monte Carlo simulation method of the present invention remains unchanged, and the maximum error is about 2.8%. It shows that the aggregation model constructed by the present invention can accurately describe the dynamic change process of the electric vehicle charging load. Figure 5Shows the base load of the distribution network. Electric vehicles are in the peak charging period between 17:00 and 20:00, and the peak charging coincides with the peak of the base load of the distribution network, which will lead to overload problems. Therefore, it is necessary to predict the adjustable capacity of electric vehicles to facilitate the corresponding adjustment plan of the distribution network.

[0122] Based on the comparison in the above different scenarios, aiming at the problem that the load of electric vehicles is difficult to accurately predict, the present invention proposes an aggregated load prediction method for electric vehicle groups to obtain the charging start time of electric vehicles. On this basis, a model combining fuzzy partition and high-order Markov is established. Using fuzzy partition can clearly define the charging and discharging states of electric vehicles, reduce the dimension of the state space, and avoid the problem of dimension mutation. The numerical example shows that the established aggregation model can accurately predict the charging and discharging power of electric vehicles, that is, it demonstrates the effectiveness and practicability of the technical solution of the present invention.

[0123] Embodiment 2

[0124] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the steps of the aggregated load prediction method for electric vehicle groups described above.

[0125] Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk (Hard Disk Drive, abbreviation: HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above types of memories.

[0126] Embodiment 3

[0127] This embodiment provides a computer device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it realizes the steps of the aggregated load prediction method for electric vehicle groups described above.

[0128] Such as Figure 6As shown, the computer device may include: at least one processor 71, such as a CPU (Central Processing Unit), at least one communication interface 73, a memory 74, and at least one communication bus 72. Among them, the communication bus 72 is used to realize the connection and communication between these components. Among them, the communication interface 73 may include a display screen and a keyboard. Optionally, the communication interface 73 may also include a standard wired interface and a wireless interface. The memory 74 may be a high-speed RAM memory (Random Access Memory, volatile random access memory), or a non-volatile memory, such as at least one disk memory. Optionally, the memory 74 may also be at least one storage device located far from the aforementioned processor 71. Among them, an application program is stored in the memory 74, and the processor 71 calls the program code stored in the memory 74 to execute any of the above method steps.

[0129] Among them, the communication bus 72 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 72 may be divided into an address bus, a data bus, a control bus, etc. For the sake of simplicity of representation, Figure 6 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0130] Among them, the memory 74 may include a volatile memory, such as a random-access memory (abbreviation: RAM); the memory may also include a non-volatile memory, such as a flash memory, a hard disk drive (abbreviation: HDD), or a solid-state drive (abbreviation: SSD); the memory 74 may also include a combination of the above types of memories.

[0131] Among them, the processor 71 may be a central processing unit (abbreviation: CPU), a network processor (abbreviation: NP), or a combination of a CPU and an NP.

[0132] Among them, the processor 71 may further include a hardware chip. The above-mentioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The above-mentioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0133] Optionally, the memory 74 is further configured to store program instructions. The processor 71 may call the program instructions to implement the electric vehicle group aggregated load prediction method of the present invention.

[0134] Those skilled in the art of the present technology should understand that the present invention may be implemented in many other specific forms without departing from the spirit and scope of the present invention. Based on the embodiments of the present invention, any changes and modifications made by those of ordinary skill in the art of the present invention according to the above disclosure shall fall within the protection scope of the claims.

Claims

1. A method for predicting the aggregate load of electric vehicle groups, characterized in that: The method comprises the following steps: S100, predicting the number of electric vehicles arriving at the charging station in the time series by Poisson distribution; S200, construct a two-layer discretization model of electric vehicle charging status based on a high-order Markov chain model; S300, calculating the estimated value of the transfer probability matrix and the coefficient of the high-order Markov chain model according to the double-layer discretization model of the electric vehicle quantity sequence and the electric vehicle charging state, and then determining the predicted value of the charging load.

2. The method for predicting the aggregate load of electric vehicle groups according to claim 1, characterized in that: S100 includes the following steps: S110, let n t is the total number of electric vehicles entering the charging station for charging services as of time t, which follows a Poisson distribution with parameters: Among them, P(n t =m) represents the probability, λ represents the number of electric vehicles arriving at the charging station per unit time; S120. Generate a sequence of electric vehicles that arrive at a charging station for charging service within the nth unit time period according to the properties of the Poisson process. The expression is as follows: Among them, K(t) is the randomly generated time series of the number of electric vehicles arriving in the total time, k n is the number of electric vehicles arriving at the charging station in the nth time slot, n is the number of segments per unit time in the total time period, T2 interval is the length of the unit time period, and T1 is the length of the total time period; S130. Assuming that the vehicle can enter the charging station without queuing and provide charging service immediately, the time when the electric vehicle arrives at the charging station at this time is the charging start time. The charging start time is calculated as follows: Where T3 is the charging start time, n' is the sequence number of the time slot, and m' is the sequence value of the electric vehicle arriving at the nth time slot; S140. Calculate the expected value X of the Poisson distribution using the following formula:

3. The method for predicting the aggregate load of electric vehicle groups according to claim 2, characterized in that: S200 includes the following steps: S210, High-order Markov description of the charging process; S220: Discretization of charging state.

4. The method for predicting the aggregate load of electric vehicle groups according to claim 3, characterized in that: S210 includes the following steps: S2101. The discrete time recursive formula for the remaining battery power is expressed as: Where S(t+1) and S(t-h+1) represent the state of charge at time t+1 and t-h+1 respectively, h represents the order, Pc represents the charging power, η represents the charging efficiency, C B is the actual capacity of the battery, Δt is the time interval; S2102. The battery state of charge is discretized into multiple state spaces, and a load process chain of each state interval based on a high-order Markov chain model is represented for dynamic conversion.

5. The method for predicting the aggregated load of electric vehicle groups according to claim 4, characterized in that: S220 includes the following steps: S2201, the battery state of charge is regarded as a fuzzy concept, the percentage of the state of charge is [0, 100], and its tertile is divided into three large intervals U1, U2, and U3, U1 is a small battery capacity state [0, 1], U2 is a normal battery capacity state [1, 80], and U3 is a large battery capacity state [80, 100]; S2202, the second level of discretization is continued on the basis of the three large intervals U1, U2, and U3, each large interval is subdivided into n small intervals, and S′(i+1) is used to represent the lower limit of the charging state of the i+1 large interval. S′(i, j) represents the lower limit value of the state of charge of the j-th small interval in the i-th large interval.

6. The method for predicting the aggregate load of electric vehicle groups according to claim 5, characterized in that: S300 includes the following steps: S310, assuming that the matrix composed of the second-order transition probability of the charging state sequence from state interval i to state interval j is: in, It represents the probability value of changing from the small battery capacity state to the small battery capacity state; Indicates the probability value of changing from a small battery capacity state to a normal battery capacity state; Indicates the probability value of changing from a small battery capacity state to a large battery capacity state; Indicates the probability value of changing from normal battery capacity state to small battery capacity state; represents the probability value of the normal battery capacity state to the normal battery capacity state; Indicates the probability value of changing from normal battery capacity state to large battery capacity state; Indicates the probability value of the state from large battery capacity to small battery capacity; Indicates the probability value of changing from a large battery capacity state to a normal battery capacity state; Indicates the probability value of the large battery capacity state to the large battery capacity state; S320、F h The estimated value of is defined as The calculation formula is as follows: in, Indicates the estimated value from the small battery capacity state to the small battery capacity state, Indicates the estimated value from the small battery capacity state to the normal battery capacity state; Indicates the estimated value from the small battery capacity state to the large battery capacity state; Indicates the estimated value from the normal battery capacity state to the small battery capacity state; represents an estimate of a normal battery capacity state to a normal battery capacity state; Indicates the estimated value from the normal battery capacity state to the large battery capacity state; Indicates the estimated value of the state from large battery capacity to small battery capacity; Indicates the estimated value from the large battery capacity state to the normal battery capacity state; Indicates a large battery capacity state to a large battery capacity state estimate; S330, let the charging state sequence converge to the Poisson distribution expected value X. The occurrence ratio of each state in the sequence is calculated as the estimated value of the Poisson distribution expected value X. Then, the coefficient λ of the high-order Markov chain model h The estimated value of is calculated by the following formula: Among them, min||·|| ∞ represents the infinite norm of the vector, λ h Satisfy two conditions: ∑λ h =1 and λ h ≥0; S340, the electric vehicle charging load value prediction model is obtained as follows: Among them, X c (t) = [X c (t,1),X c (t,2),X c (t,3)] T is a 3D column vector representing the load value at time t, X c (t+1) represents the predicted value of the electric vehicle load at the charging station at time t+1, L is the maximum order of the Markov chain, where L=2, X c (t-h+1) represents the charging load of electric vehicles in the charging station at time t-h+1, Y c (t) represents the output power of the electric vehicle at time t, C c is the output matrix, which is a 3D unit row vector, V c (t) is the load change caused by external factors and is related to X c (t) has the same structure, P′ max is the maximum charging power; in, and is the number of electric vehicles plugged in and unplugged in the tth time interval, N ev (t) is the total number of electric vehicles in the charging station at time t, X in (t) is the charging power at the tth time interval, X out (t) is the discharge power at the tth time interval.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the electric vehicle group aggregate load prediction method as described in any one of claims 1 to 6 are implemented.

8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the electric vehicle group aggregate load prediction method as described in any one of claims 1-6 are implemented.

Images