CN-121998251-A - Sobol sensitivity analysis method and system for electric vehicle charging station

CN121998251ACN 121998251 ACN121998251 ACN 121998251ACN-121998251-A

Abstract

The invention provides a Sobol sensitivity analysis method and a Sobol sensitivity analysis system for an electric vehicle charging station, which relate to the technical field of external influence factor analysis of adjustable load of the electric vehicle charging station and comprise the following steps of establishing user charging decision models of multiple types of electric vehicle users; the method comprises the steps of carrying out global sensitivity quantitative analysis on a plurality of key influence parameters of a charging station, generating a parameter interaction effect matrix by identifying and quantifying nonlinear coupling relations among the key influence parameters based on the result of the global sensitivity quantitative analysis, establishing a standardized mathematical model of multiple types of extreme events, substituting a user charging decision model into the standardized mathematical model, and carrying out comprehensive risk assessment through a multistage pressure test method and a toughness score calculation model based on the parameter interaction effect matrix. The method has the advantages that accurate parameter interaction effect accurate identification, comprehensive evaluation of extreme scene risks and quantitative evaluation of system toughness are realized.

Inventors

YANG YING
SONG JINGYI
LI HAORAN
LI YULIN

Assignees

清华四川能源互联网研究院

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (10)

1. The Sobol sensitivity analysis method for the electric vehicle charging station is characterized by comprising the following steps of: Establishing a user charging decision model of multiple types of electric automobile users, wherein the types of the electric automobile users comprise commuting users with periodic charging behaviors, flexible users for determining charging time based on electricity price change, night users charged in night valley time and random users with irregular charging time; Performing global sensitivity quantization analysis on a plurality of key influence parameters of the charging station, including constructing a basic sampling matrix A, an independent sampling matrix B and a parameter disturbance matrix of an ith key influence parameter , N is the dimension of key influence parameters of the charging station, and the first-order main effect of each key influence parameter is calculated; Generating a parameter interaction effect matrix by identifying and quantifying a nonlinear coupling relationship between key influence parameters based on the result of the global sensitivity quantitative analysis; establishing a standardized mathematical model of multiple types of extreme events, substituting a user charging decision model into the standardized mathematical model, and acquiring load distribution statistics under a multipole scene; And based on the parameter interaction effect matrix, carrying out comprehensive risk assessment through a multi-stage pressure test method and a toughness fraction calculation model.
2. The Sobol sensitivity analysis method for an electric vehicle charging station of claim 1, wherein the user charging decision model comprises: the charging time t of the commute user decision model obeys a normal distribution with a mean value of 72 and a standard deviation of 4: ; wherein, the charging time is represented by a number of time units, and each 15 minutes is a time unit; The flexible user decision model is built based on real-time electricity prices: ; Wherein, the For real-time electricity prices (yuan/kWh), Is a charge response coefficient; The charging time t of the night user decision model obeys the mean distribution of intervals 4 to 24: ; The charging time t obeying mean value of the random user decision model is Standard deviation is Normal distribution of (c): ; Wherein, the And All from system parameters.
3. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 1, wherein the method of global sensitivity quantization analysis of a plurality of key influence parameters of the charging station is: Generating basic sampling points by adopting an M multiplied by N dimension Sobol low-difference sequence, wherein M is the number of samples, and obtaining a basic sampling matrix A; generating a quasi-random sampling matrix with the same dimension as the basic sampling matrix A but completely independent sequence to obtain an independent sampling matrix B; replacing the ith column of the independent sampling matrix B with the ith column of the basic sampling matrix A to obtain the parameter disturbance matrix of the ith key influence parameter , ; Scaling the standard unit interval [0,1] of the Sobol sequence to the actual value range of the key influence parameter through a linear transformation function; and calculating a Sobol sensitivity index.
4. A Sobol sensitivity analysis method for an electric vehicle charging station according to claim 3, wherein the method of mapping the standard unit interval [0,1] of the Sobol sequence to the actual key influence parameter value range is: ; Wherein, the Is a numerical value mapped to the actual key influence parameter value range, And Respectively minimum and maximum values of the corresponding key influencing parameters, Sampling values of the corresponding key influence parameters in the Sobol low-difference sequences; the method for calculating the sensitivity index of the Sobol comprises the following steps: first order main effect of obtaining ith key influence parameter : ; Obtaining the total effect : ; Wherein, the 、、 Respectively a basic sampling matrix A, an independent sampling matrix B and a parameter disturbance matrix Is a charge load model output vector of (a), As the average value of all the outputs, To output the total variance.
5. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 1, wherein the method of generating a parameter interaction effect matrix is: Respectively replacing the ith column and the jth column of the independent sampling matrix B with the ith column and the jth column of the basic sampling matrix A to obtain a dual-parameter interaction matrix of the ith key influence parameter and the jth key influence parameter , , , ; Calculating the second order interaction effect of the ith key influence parameter and the jth key influence parameter : ; Wherein, the And Respectively the basic sampling matrix And the dual-parameter interaction matrix The corresponding charge load model outputs a vector, And First order main effects of the ith key influence parameter and the jth key influence parameter respectively; Traversing to calculate all Later, will And placing the ith row and the jth column of the parameter interaction effect matrix to form the parameter interaction effect matrix.
6. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 1, wherein the method of establishing a standardized mathematical model of a plurality of classes of extreme events is: setting a plurality of extreme scenes, and respectively modeling each extreme scene, wherein the extreme scenes comprise a concentrated charging extreme scene, a power grid constraint extreme scene, an extreme weather scene, an equipment fault extreme scene, a demand surge extreme scene and an electricity price impact extreme scene; The method for acquiring the load distribution statistics under the multipole scene is to respectively perform Monte Carlo extremum statistical analysis aiming at each extreme scene.
7. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 6, wherein the method of modeling each extreme scene separately includes: Modeling of concentrated charging extreme scenes by setting simultaneous charging proportionality coefficients And a time concentration factor Describing the centralized charging extreme scenario; Modeling of power grid constraint extreme scene, and setting power grid capacity reduction coefficient And line impedance growth coefficient Describing the grid constraint extreme scene; modeling extreme weather scenes by setting ambient temperature offsets And corresponding vehicle energy consumption growth coefficient Describing the extreme weather scenario; modeling of equipment fault extreme scenes, and setting initial random fault rate of charging equipment And cascade failure trigger probability Describing the equipment fault extreme scene; modeling of extreme scene of demand proliferation, and setting of quantity proliferation coefficient of electric automobile And average mileage increase coefficient of bicycle Describing the demand surge extreme scenario Modeling of electricity price impact extreme scenes by setting price multiplication coefficients of electric power markets And a user behavior response delay parameter describing the electricity price impact extreme scene.
8. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 7, wherein the method of separately performing the monte carlo extremum statistical analysis for each of the extreme scenes is to separately perform the following operations for each of the extreme scenes: Carrying out random sampling for a plurality of times in the extreme scene, and carrying out sampling operation on user charging decision models of various electric automobile users each time to obtain a system peak load sample; and counting to obtain the average value, standard deviation, 95% quantile value, 99% quantile value, 99.9% quantile value and observation maximum value of the load distribution, and obtaining the load distribution statistics under the multipole scene.
9. The Sobol sensitivity analysis method for an electric vehicle charging station according to claim 1, wherein the method for comprehensive risk assessment by a multistage pressure test method and a toughness score calculation model is as follows: Setting four standardized pressure test scenes of different intensities, including: double load test, wherein the load multiplication coefficient is 2.0, and the duration is 4 hours, and the double load test is used for simulating daily operation peak pressure; three times of load test, wherein the load multiplication coefficient is 3.0, and the duration is 2 hours, and the test is used for simulating short-term impact pressure; continuously testing high load, wherein the load multiplication coefficient is 1.8, and the duration is 8 hours, so as to simulate long-term heavy-load operating pressure; the extreme peak test, the load multiplication coefficient is 4.0, the duration is 1 hour, is used for simulating the ultimate pressure of the bearing capacity of the system; Aiming at each standardized pressure test scene with different intensity, the toughness score calculation model is established through a comprehensive toughness evaluation function respectively: ; wherein R is the toughness fraction, As a basic toughness coefficient, the basic toughness coefficient is determined according to the system bearing capacity of standardized pressure test scenes with different intensities, In order to penalize the coefficient of violation, In order to constrain the rate of the violation, Is a self-adaptive adjusting coefficient, the value range is 0.8 to 1.2, and min is a function of the minimum value.
10. Sobol sensitivity analysis system for an electric vehicle charging station, applied to a Sobol sensitivity analysis method for an electric vehicle charging station according to any one of claims 1 to 9, characterized by comprising: The user behavior modeling module is used for establishing user charging decision models of various electric automobile users, wherein the types of the electric automobile users comprise commuting users with periodic charging behaviors, flexible users for determining charging time based on electricity price change, night users charged in night valley time periods and random users with irregular charging time; the user behavior module is used for carrying out global sensitivity quantitative analysis on a plurality of key influence parameters of the charging station, and comprises the steps of constructing a basic sampling matrix A, an independent sampling matrix B and a parameter disturbance matrix of an ith key influence parameter , N is the dimension of key influence parameters of the charging station, and the first-order main effect of each key influence parameter is calculated; The parameter interaction analysis module is used for generating a parameter interaction effect matrix by identifying and quantifying a nonlinear coupling relation between key influence parameters based on the result of the global sensitivity quantitative analysis; The extreme scene generation module is used for establishing standardized mathematical models of multiple types of extreme events, and then substituting the user charging decision model into the standardized mathematical model to acquire load distribution statistics under the multipole scene; The system toughness evaluation module is used for carrying out comprehensive risk evaluation through a multistage pressure test method and a toughness fraction calculation model based on the parameter interaction effect matrix.

Description

Sobol sensitivity analysis method and system for electric vehicle charging station Technical Field The invention relates to the technical field of analysis of external influence factors of adjustable loads of electric vehicle charging stations, in particular to a Sobol sensitivity analysis method and system for electric vehicle charging stations. Background The electrical load of an electric vehicle charging station is not only influenced by the rated parameters of the device, but also depends significantly on external factors. By carrying out system analysis and modeling on external influence factors, a predictable and controllable load change rule under the influence of the external factors can be extracted, and a reliable basis is provided for power grid dispatching and planning. The existing method for analyzing external influence factors has some defects, such as that the method based on single factor sensitivity analysis can only reflect local influence of parameters and cannot capture interaction among parameters, the method based on regression analysis can handle parameter interaction but has limited applicability to a highly nonlinear system, the parameter importance sorting based on Morris screening method is suitable for preliminary parameter screening but has relatively limited precision, the extreme scene analysis of the traditional Monte Carlo method lacks systematic extreme event classification and standardized modeling, and is difficult to comprehensively evaluate influence of different types of extreme scenes, and the power grid risk evaluation based on scene analysis. That is, in the prior art, aiming at the traditional load setting, the particularity of the electric vehicle charging station as the adjustable load is not fully considered, the performance of the system under various extreme conditions cannot be comprehensively evaluated, and complex interaction relations among a plurality of influencing factors, particularly coupling effects of factors such as temperature, electricity price, user behavior and the like, are difficult to accurately identify and quantify, so that the understanding of the system behavior is not deep enough, quantitative evaluation indexes such as system toughness, risk level and the like are lacking, and accurate numerical basis is difficult to provide for decision making. Therefore, the analysis of external influence factors aiming at the adjustable load of the electric vehicle charging station needs to be optimized, so that accurate parameter interaction effect accurate identification, comprehensive evaluation of extreme scene risk and quantitative evaluation of system toughness are realized. Disclosure of Invention The invention aims to provide a Sobol sensitivity analysis method and a system for an electric vehicle charging station, which realize accurate parameter interaction effect identification, comprehensive evaluation of extreme scene risk and quantitative evaluation of system toughness. The invention is realized by the following technical scheme: the Sobol sensitivity analysis method for the electric vehicle charging station comprises the following steps: Establishing a user charging decision model of multiple types of electric automobile users, wherein the types of the electric automobile users comprise commuting users with periodic charging behaviors, flexible users for determining charging time based on electricity price change, night users charged in night valley time and random users with irregular charging time; Performing global sensitivity quantization analysis on a plurality of key influence parameters of the charging station, including constructing a basic sampling matrix A, an independent sampling matrix B and a parameter disturbance matrix of an ith key influence parameter ,N is the dimension of key influence parameters of the charging station, and the first-order main effect of each key influence parameter is calculated; Generating a parameter interaction effect matrix by identifying and quantifying a nonlinear coupling relationship between key influence parameters based on the result of the global sensitivity quantitative analysis; establishing a standardized mathematical model of multiple types of extreme events, substituting a user charging decision model into the standardized mathematical model, and acquiring load distribution statistics under a multipole scene; And based on the parameter interaction effect matrix, carrying out comprehensive risk assessment through a multi-stage pressure test method and a toughness fraction calculation model. Preferably, the user charging decision model comprises: the charging time t of the commute user decision model obeys a normal distribution with a mean value of 72 and a standard deviation of 4: ; wherein, the charging time is represented by a number of time units, and each 15 minutes is a time unit; The flexible user decision model is built based on real-time electricity prices: ; Wherein, the