Charging pile load prediction method based on improved bandwidth, storage medium and equipment

Abstract

The invention relates to the technical field of load prediction and management of electric vehicle charging piles, in particular to a charging pile load prediction method based on bandwidth improvement, a storage medium and equipment. The method comprises the steps of obtaining a standardized data set; substituting the initial bandwidth h1 into a kernel density estimation function to obtain an initial probability density # imgabs0 # for load prediction of the charging pile, performing m-1 times of iterative optimization on the # imgabs1 # to obtain an updated probability density # imgabs2 #, and calculating an integral mean square error (MISE); when the MISE is smaller than Eth, iterative optimization is stopped, and the optimal bandwidth h * m is obtained; substituting h * m into a kernel density estimation function to obtain an optimal probability density # imgabs3 #, wherein a computer program instruction is stored on the storage medium; the apparatus includes computer program instructions stored in a memory; the method is implemented when computer program instructions are executed by a processor. According to the invention, the problem of large prediction error caused by inaccurate bandwidth parameter selection is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of load prediction and management of electric vehicle charging piles, and in particular to a charging pile load prediction method based on improved bandwidth, a storage medium and a device. Background Art

[0002] With the popularity of electric vehicles, the construction and management of charging infrastructure has become one of the key guarantees for the widespread use of electric vehicles. As the main facility for charging electric vehicles, the load of charging piles directly affects the load management of the power grid and the operating efficiency of charging piles. However, since the charging demand of electric vehicles is affected by many factors, including vehicle type, driving habits, charging time, climatic conditions and number of users, the load of charging piles is highly uncertain and volatile. How to accurately predict the load demand of charging piles and conduct effective resource scheduling has become a major problem in charging pile management.

[0003] Most of the existing charging pile load prediction methods are based on traditional time series analysis or machine learning models, but these methods often have the following shortcomings: 1) Poor model generalization ability: Traditional prediction models are usually based on fixed assumptions and parameters, and it is difficult to adapt to the complexity of charging pile load data. 2) Bandwidth parameter selection usually relies on fixed or empirical bandwidth selection strategies: When using non-parametric methods such as kernel density estimation, the choice of bandwidth parameters has a crucial impact on prediction accuracy; traditional methods usually rely on fixed or empirical bandwidth selection strategies, which cannot adapt to local fluctuations in charging pile load data and may lead to large prediction errors. 3) Machine learning models require a large amount of sample data, and the training process takes a lot of time and computing power, and the accuracy is not high.

[0004] In summary, it is necessary to provide a charging pile load prediction method, storage medium and device based on improved bandwidth to solve the problem of large prediction error caused by inaccurate bandwidth parameter selection in existing charging pile load prediction methods. Summary of the invention

[0005] The present invention aims to provide a charging pile load prediction method, storage medium and device based on improved bandwidth. The specific technical solution is as follows:

[0006] In a first aspect, the present invention provides a charging pile load prediction method based on improved bandwidth, comprising:

[0007] Step S1, obtaining a charging pile load sample data set, and preprocessing it to obtain a standardized data set;

[0008] Step S2: Calculate the total number of samples N and the variance σ of the charging pile load samples in the standardized data set, and combine the improved bandwidth function to obtain the initial bandwidth h1; substitute the initial bandwidth h1 into the kernel density estimation function to obtain the initial probability density of the charging pile load prediction.

[0009] Step S3: The initial probability density Perform m-1 iterations of optimization to obtain the updated probability density Wherein, m is the number of iterations, which is an integer greater than or equal to 2;

[0010] Step S4: Calculation The integrated mean square error MISE of the sample histogram distribution of the standardized data set; the MISE is compared with the error threshold E th Compare; if the MISE is less than E th , then the iterative optimization stops, and the current corresponding bandwidth is the optimal bandwidth h * m ; If the MISE is greater than or equal to E th , then continue to iterate and optimize according to step S3 until the MISE is less than E th , get the optimal bandwidth h * m ; Set the optimal bandwidth h * m Substitute into the kernel density estimation function to obtain the optimal probability density

[0011] Optionally, in step S2, the improved bandwidth function is expressed by formula (1):

[0012]

[0013] When the initial bandwidth h1 is calculated using formula (1), the values ​​of c3 and c4 are both zero; h represents the improved bandwidth function.

[0014] Optionally, in step S2, the kernel density estimation function is expressed by formula (2):

[0015]

[0016] in, is the kernel density estimation function; is the Gaussian kernel function; x is the charging pile load sample; x t is the tth charging pile load sample; t is an integer greater than or equal to 1 and less than or equal to N.

[0017] Optionally, in step S3, the iterative optimization process includes:

[0018] Step S3.1: The initial probability density Perform Fourier transform to obtain the characteristic function; the characteristic function is expressed by formula (3):

[0019]

[0020] In formula (3), when m is 2, for Express Fourier transform; F x (ω) represents the characteristic function; i is the imaginary part in mathematics; ω represents the angular frequency; μ represents the mean value of the charging pile load sample; c 3,m represents the optimal bandwidth parameter of the third-order cumulant when the number of iterations is m; c 4,m It represents the optimal bandwidth parameter of the fourth-order cumulant when the number of iterations is m;

[0021] Step S3.2, taking the real part of the characteristic function to obtain a real part function; the real part function is expressed by formula (4):

[0022]

[0023] In formula (4), Express After Fourier transformation, take the real part of the first four orders;

[0024] The imaginary part of the characteristic function is taken to obtain the imaginary part function; the imaginary part function is expressed by formula (5):

[0025]

[0026] In formula (5), Express After Fourier transformation, take the imaginary part of the first four orders;

[0027] Step S3.3, solving the real part function and the imaginary part function together to obtain c 3,m and c 4,m ; c 3,m and c 4,m Substitute into the improved bandwidth function in step S2 to obtain the iteratively updated bandwidth h m ; The iteratively updated bandwidth h m Substitute into the kernel density estimation function to obtain the iteratively updated probability density

[0028] Optionally, in step S4, the MISE is calculated using formula (6):

[0029]

[0030] In formula (6), X n is the number of samples falling into the nth interval of the charging pile load sample; b is the interval width; S is the number of intervals of the histogram distribution of the standardized data set samples; n is an integer greater than or equal to 1 and less than or equal to S; Indicates that in the interval b(n-1) to b·n Perform integration.

[0031] Optionally, in step S4, the error threshold E th The value of is less than or equal to 0.985.

[0032] Optionally, in step S1, the preprocessing includes sequentially performing data cleaning and standardization processing on the acquired charging pile load sample data set.

[0033] Optionally, the data cleaning includes removing outliers and removing missing values.

[0034] In a second aspect, the present invention provides a computer storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the charging pile load prediction method based on improved bandwidth.

[0035] In a third aspect, the present invention provides an electronic device comprising: at least one processor, at least one memory, and computer program instructions stored in the memory, which implement the charging pile load prediction method based on improved bandwidth when the computer program instructions are executed by the processor.

[0036] The application of the technical solution of the present invention has at least the following beneficial effects:

[0037] The present invention provides a charging pile load prediction method, storage medium and device based on improved bandwidth, which can quickly and accurately solve the optimal bandwidth h * m , while improving the estimation accuracy, significantly reducing the operation time, thereby improving the accuracy and stability of charging pile load prediction, better capturing the changing rules in load data, reducing noise interference, and being able to provide more accurate load prediction results in uncertain environments. Therefore, the present invention can accurately predict the load demand of charging piles, provide a guarantee for power grid operators to better predict charging loads and conduct power grid dispatching, avoid power overload or waste, and solve the problem of large prediction errors caused by inaccurate bandwidth parameter selection in existing charging pile load prediction methods.

[0038] In addition to the above-described purposes, features and advantages, the present invention has other purposes, features and advantages. The present invention will be further described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings constituting a part of this application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0040] Figure 1 It is a flow chart of a charging pile load prediction method based on improved bandwidth in an embodiment. DETAILED DESCRIPTION

[0041] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0042] Example:

[0043] See also Figure 1 , a charging pile load prediction method based on improved bandwidth includes:

[0044] Step S1, obtain a charging pile load sample data set (the data set comes from a free data set provided by the Kaggle website https: / / www.kaggle.com, specifically the charging load statistics of a charging pile in a certain place, including 29600 days of charging load charging power and connection time statistics of a charging pile in a certain place, and the charging load power unit is kW), and obtain a standardized data set after preprocessing;

[0045] Step S2: Calculate the total number of samples N and the variance σ of the charging pile load samples in the standardized data set, and combine the improved bandwidth function to obtain the initial bandwidth h1; substitute the initial bandwidth h1 into the kernel density estimation function to obtain the initial probability density of the charging pile load prediction.

[0046] Step S3: The initial probability density Perform m-1 iterations of optimization to obtain the updated probability density Wherein, m is the number of iterations, which is an integer greater than or equal to 2;

[0047] Step S4: Calculation The integrated mean square error MISE of the sample histogram distribution of the standardized data set; the MISE is compared with the error threshold E th Compare; if the MISE is less than E th , then the iterative optimization stops, and the current corresponding bandwidth is the optimal bandwidth h *m ; If the MISE is greater than or equal to E th , then continue to iterate and optimize according to step S3 until the MISE is less than E th , get the optimal bandwidth h * m ; Set the optimal bandwidth h * m Substitute into the kernel density estimation function to obtain the optimal probability density

[0048] In step S2, the improved bandwidth function is expressed by formula (1):

[0049]

[0050] When the initial bandwidth h1 is calculated using formula (1), the values ​​of c3 and c4 are both zero; h represents the improved bandwidth function.

[0051] In step S2, the kernel density estimation function is expressed by formula (2):

[0052]

[0053] in, is the kernel density estimation function; is the Gaussian kernel function; x is the charging pile load sample; x t is the tth charging pile load sample; t is an integer greater than or equal to 1 and less than or equal to N.

[0054] In step S3, the iterative optimization process includes:

[0055] Step S3.1: The initial probability density Perform Fourier transform to obtain the characteristic function; the characteristic function is expressed by formula (3):

[0056]

[0057] In formula (3), when m is 2, for Express Fourier transform; F x (ω) represents the characteristic function; i is the imaginary part in mathematics; ω represents the angular frequency; μ represents the mean value of the charging pile load sample; c 3,m represents the optimal bandwidth parameter of the third-order cumulant when the number of iterations is m; c 4,m It represents the optimal bandwidth parameter of the fourth-order cumulant when the number of iterations is m;

[0058] Step S3.2, taking the real part of the characteristic function to obtain a real part function; the real part function is expressed by formula (4):

[0059]

[0060] In formula (4), Express After Fourier transformation, take the real part of the first four orders;

[0061] The imaginary part of the characteristic function is taken to obtain the imaginary part function; the imaginary part function is expressed by formula (5):

[0062]

[0063] In formula (5), Express After Fourier transformation, take the imaginary part of the first four orders;

[0064] Step S3.3, solving the real part function and the imaginary part function together to obtain c 3,m and c 4,m ; c 3,m and c 4,m Substitute into the improved bandwidth function in step S2 to obtain the iteratively updated bandwidth h m ; The iteratively updated bandwidth h m Substitute into the kernel density estimation function to obtain the iteratively updated probability density

[0065] In step S4, the MISE is calculated using formula (6):

[0066]

[0067] In formula (6), X n is the number of samples falling into the nth interval of the charging pile load sample; b is the interval width; S is the number of intervals of the histogram distribution of the standardized data set samples; n is an integer greater than or equal to 1 and less than or equal to S; Indicates that in the interval b(n-1) to b·n Perform integration

[0068] In step S4, the error threshold E th The value of is less than or equal to 0.985.

[0069] In step S1, the preprocessing includes sequentially performing data cleaning and standardization on the acquired charging pile load sample data set. The data cleaning includes removing outliers and removing missing values.

[0070] The charging pile load prediction method based on improved bandwidth (abbreviated as GC-KDE), empirical estimation method (abbreviated as NR-KDE), interpolation method (abbreviated as PI-KDE), genetic algorithm kernel density estimation method (abbreviated as GA-KDE) and particle swarm algorithm kernel density estimation method (abbreviated as PSO-KDE) described in this embodiment are used to process the standardized data set in this embodiment to obtain MISE data and the optimal probability density of the charging pile load. Estimate the time required (in seconds). MISE data is shown in Table 1; the optimal probability density of charging pile load See Table 2 for an estimate of the time required.

[0071] Table 1 MISE data

[0072]

[0073] Table 2 Optimal probability density of charging pile load Estimated time required

[0074]

[0075] From the data in Tables 1 and 2, it can be seen that compared with the empirical estimation method (abbreviated as NR-KDE), the interpolation method (abbreviated as PI-KDE), the genetic algorithm kernel density estimation method (abbreviated as GA-KDE) and the particle swarm algorithm kernel density estimation method (abbreviated as PSO-KDE), the charging pile load prediction method based on improved bandwidth (abbreviated as GC-KDE) in this embodiment has the smallest MISE data and the shortest running time; wherein the running time refers to the optimal probability density of the load charging power. Estimate the time required and the optimal probability density for charging duration Estimate the time required.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A charging pile load prediction method based on improved bandwidth, characterized in that: include: Step S1, obtaining a charging pile load sample data set, and preprocessing it to obtain a standardized data set; Step S2: Calculate the total number of samples N and the variance σ of the charging pile load samples in the standardized data set, and combine the improved bandwidth function to obtain the initial bandwidth h1; substitute the initial bandwidth h1 into the kernel density estimation function to obtain the initial probability density of the charging pile load prediction. Step S3: The initial probability density Perform m-1 iterations of optimization to obtain the updated probability density Wherein, m is the number of iterations, which is an integer greater than or equal to 2; Step S4: Calculation The integrated mean square error MISE of the sample histogram distribution of the standardized data set; the MISE is compared with the error threshold E th Compare; if the MISE is less than E th , then the iterative optimization stops, and the current corresponding bandwidth is the optimal bandwidth h * m ; If the MISE is greater than or equal to E th , then continue to iterate and optimize according to step S3 until the MISE is less than E th , get the optimal bandwidth h * m ; Set the optimal bandwidth h * m Substitute into the kernel density estimation function to obtain the optimal probability density 2. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S2, the improved bandwidth function is expressed by formula (1): When the initial bandwidth h1 is calculated using formula (1), the values ​​of c3 and c4 are both zero; h represents the improved bandwidth function.

3. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S2, the kernel density estimation function is expressed by formula (2): in, is the kernel density estimation function; is the Gaussian kernel function; x is the charging pile load sample; x t is the tth charging pile load sample; t is an integer greater than or equal to 1 and less than or equal to N.

4. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S3, the iterative optimization process includes: Step S3.1: The initial probability density Perform Fourier transform to obtain the characteristic function; the characteristic function is expressed by formula (3): In formula (3), when m is 2, for Express Fourier transform; F x (ω) represents the characteristic function; i is the imaginary part in mathematics; ω represents the angular frequency; μ represents the mean value of the charging pile load sample; c 3,m represents the optimal bandwidth parameter of the third-order cumulant when the number of iterations is m; c 4,m It represents the optimal bandwidth parameter of the fourth-order cumulant when the number of iterations is m; Step S3.2, taking the real part of the characteristic function to obtain a real part function; the real part function is expressed by formula (4): In formula (4), Express After Fourier transformation, take the real part of the first four orders; The imaginary part of the characteristic function is taken to obtain the imaginary part function; the imaginary part function is expressed by formula (5): In formula (5), Express After Fourier transformation, take the imaginary part of the first four orders; Step S3.3, solving the real part function and the imaginary part function together to obtain c 3,m and c 4,m ; c 3,m and c 4,m Substitute into the improved bandwidth function in step S2 to obtain the iteratively updated bandwidth h m ; The iteratively updated bandwidth h m Substitute into the kernel density estimation function to obtain the iteratively updated probability density 5. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S4, the MISE is calculated using formula (6): In formula (6), X n is the number of samples falling into the nth interval of the charging pile load sample; b is the interval width; S is the number of intervals of the histogram distribution of the standardized data set samples; n is an integer greater than or equal to 1 and less than or equal to S; Indicates that in the interval b(n-1) to b·n Perform integration.

6. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S4, the error threshold E th The value of is less than or equal to 0.

985.

7. The charging pile load prediction method based on improved bandwidth according to claim 1 is characterized in that: In step S1, the preprocessing includes sequentially performing data cleaning and standardization processing on the acquired charging pile load sample data set.

8. The charging pile load prediction method based on improved bandwidth according to claim 7 is characterized in that: The data cleaning includes removing outliers and removing missing values.

9. A computer storage medium, characterized in that Computer program instructions are stored thereon, and when the computer program instructions are executed by a processor, a charging pile load prediction method based on improved bandwidth as described in any one of claims 1 to 8 is implemented.

10. An electronic device, characterized in that: include: At least one processor, at least one memory, and computer program instructions stored in the memory, when the computer program instructions are executed by the processor, implement the charging pile load prediction method based on improved bandwidth as described in any one of claims 1 to 8.

Images