Electric vehicle charging and discharging state quantitative evaluation method based on simulation algorithm

Abstract

The invention discloses an electric vehicle charging and discharging state quantitative evaluation method based on a simulation algorithm. The method comprises the steps that a power change model of an electric vehicle in the charging and discharging process is constructed; acquiring historical charging and discharging data of each electric vehicle and the corresponding charging pile, wherein the historical charging and discharging data comprises power change calculation parameters; utilizing a Monte Carlo method to predict a daily total power change curve of a specified date for multiple times, wherein the daily total power change curve is obtained by calculating and summarizing randomly extracted historical charging and discharging data of the electric vehicle according to a power change model; and analyzing the plurality of daily total power change curves by using a clustering algorithm to obtain the daily total power change curve with the highest probability. The method is used for implementing a multi-dimensional adjustment strategy integrating photovoltaic, energy storage and charging functions in the later period.

Description

Technical Field

[0001] The present invention relates to the technical field of power consumption planning, and particularly to a method for quantitatively evaluating the charging and discharging states of electric vehicles based on a simulation algorithm. Background Art

[0002] Due to the continuous decline in the cost of photovoltaic construction, the scale of photovoltaic power generation in rural areas, suburban areas and other regions has been continuously expanding. However, during the high-generation period, the power generated cannot be self-consumed, resulting in reverse power transmission in the transformer substation area, which affects the safe and stable operation of the power grid. On the other hand, during peak power consumption periods, the load demand increases significantly, and the power grid faces huge pressure in power transmission. Especially during summer holidays or winter heating periods, the power supply capacity of the power grid is severely tested.

[0003] In order to efficiently address the above-mentioned challenges, power grid companies have adopted a new energy electric vehicle as the basic platform, adjusted the charging and discharging behaviors through a price incentive mechanism, and implemented a multi-dimensional adjustment strategy integrating photovoltaic, energy storage and charging functions. However, during the actual use of this strategy, due to the inability to effectively grasp the changes in the charging and discharging power of electric vehicles within a certain period in the future, the final power consumption and power supply are insufficient, and the strategy results are not as expected. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for quantitatively evaluating the charging and discharging states of electric vehicles based on a simulation algorithm to solve the problem in the background art that the changes in the charging and discharging power of electric vehicles cannot be predicted currently.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A method for quantitatively evaluating the charging and discharging states of electric vehicles based on a simulation algorithm, the method comprising:

[0007] Constructing a power change model during the charging and discharging process of electric vehicles;

[0008] Obtaining the historical charging and discharging data of each electric vehicle and the corresponding charging pile, including power change calculation parameters;

[0009] Using the Monte Carlo method to predict the daily total power change curve of a specified date multiple times, where the daily total power change curve is obtained by calculating and summarizing the historical charging and discharging data of randomly selected electric vehicles according to the power change model;

[0010] Analyzing multiple daily total power change curves using a clustering algorithm to obtain the daily total power change curve with the highest probability.

[0011] Preferably, the charging power calculation model of the electric vehicle is expressed as:

[0012]

[0013] Wherein, P(t) represents the charging power at time t; P max represents the peak value of the charging power; k is a parameter determining the power growth rate; t max represents the time required for the charging power to reach the peak value, t fall represents the time from which the charging power max drops to 0, t full represents the charging end time.

[0014] Preferably, the discharge power calculation model of the electric vehicle is expressed as:

[0015]

[0016] Wherein, P maxrel is the maximum discharge power of the electric vehicle battery, t ramp is the time for the discharge power to increase from 0 to P maxrel required, t decay is the time for the discharge power to decrease from P maxrel to 0, t end is the discharge end time.

[0017] Preferably, the use of the Monte Carlo method to predict the charging and discharging power of electric vehicles on a specified date multiple times includes:

[0018] Initial setting step, setting calculation parameters, including the total number of cycles M and the total number of electric vehicles N;

[0019] Random extraction step, randomly extracting the status data of N electric vehicles from historical data, obtaining the charging and discharging status time periods of electric vehicles in one day, combining the data when the electric vehicle is connected to the charging pile, and using the power change model to obtain the power change curves of each electric vehicle during charging and discharging, and superimposing and calculating the charging and discharging power change curves of all electric vehicles to obtain the daily total power change curve;

[0020] Circulation step, repeating the random extraction step M times in total to obtain M daily total power change curves.

[0021] Preferably, in the Monte Carlo method for predicting the charging and discharging power of electric vehicles on a specified date multiple times, a convergence determination step is further included: after obtaining M daily total power change curves, taking out the charging power change curves of electric vehicles in each region to calculate the charging power proportion in each region, analyzing the supply-demand difference using the charging power proportion and charging demand proportion in each region, and determining convergence when the supply-demand difference is less than a preset threshold and shows a decreasing trend, otherwise repeating the random extraction step and the circulation step.

[0022] Preferably, the supply-demand difference is accumulated from the difference value between the charging power proportion and the charging demand proportion in each region.

[0023] Preferably, the charging power ratio of each region is expressed as:

[0024]

[0025] In the formula, is the average charging power, i represents the current region, N is the total number of regions, and j is the region index.

[0026] Preferably, in the charging demand ratio, the charging demand is characterized by the number of charging piles in each region. For each region, the ratio of its charging demand to the total demand is calculated and expressed as:

[0027]

[0028] Among them, U i represents the charging demand ratio of the i-th region, Q i represents the number of charging piles in the i-th region, and N is the total number of regions.

[0029] Preferably, in the clustering algorithm, all daily charge-discharge total power change curves are classified, the dominant category of the curves is determined according to the majority principle, and the daily charge-discharge total power change curves under the dominant category are used for data fitting to obtain the ideal daily charge-discharge power change curve of electric vehicles.

[0030] Preferably, in the data fitting, after conversion with the center point of the dominant category, it is used as the daily total power change curve of electric vehicles with the highest probability.

[0031] Compared with the prior art, the beneficial effects of the present invention are:

[0032] The present invention uses the Monte Carlo algorithm to perform a large number of simulations on the charging and discharging power of electric vehicles, predicts the daily total power curve of electric vehicles on a specified date multiple times, and performs clustering analysis on the daily total power curves of electric vehicles on all specified dates to extract the most likely daily total power curve for the subsequent implementation of multi-dimensional adjustment strategies integrating photovoltaic, energy storage, and charging functions. Description of the Drawings

[0033] Figure 1 It is a flowchart of a method for quantitatively evaluating the charging and discharging state of an electric vehicle provided in this embodiment.

[0034] Figure 2 It is a flowchart of using the Monte Carlo method to predict the daily total power change curve on a specified date multiple times. Detailed Embodiment

[0035] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] A method for quantitatively evaluating the charging and discharging states of an electric vehicle based on a simulation algorithm, the method comprising the following steps:

[0037] Step 1, respectively construct power change models of an electric vehicle during the charging process and the discharging process.

[0038] (1) Analysis of the power change of the electric vehicle during the charging process:

[0039] Since the charging power is related to the charging method, and the current charging methods mainly include constant current, stage, pulse charging, etc. In this embodiment, for simplicity of analysis, it is assumed that the power will reach the maximum within a short time at the beginning of charging and will decrease to 0 at a constant rate during the subsequent charging process. Specifically:

[0040] In the initial stage of charging, the power of the charging pile rises rapidly to a maximum value. This stage can be expressed by the following formula:

[0041] P(t) = P max (1 - e -kt ), 0 < t < t max

[0042] In the formula, P(t) represents the charging power at time t; P max represents the peak value of the charging power; k is a parameter that determines the power growth rate; t max represents the time required for the charging power to reach the peak value.

[0043] During the charging process stage, the charging power remains at the peak value for a period of time. This stage can be expressed by the following formula:

[0044] P(t) = P max , t max ≤ t < t full - t fall In the formula, t fall represents the time required for the charging power to drop from t max to 0, and t full represents the charging end time.

[0045] In the end stage of charging, the charging power gradually decreases after reaching the peak value until the charging power P(t) equals 0. This stage can be expressed by the following formula:

[0046] P(t) = Pmax -at, where t full -t fall ≤t≤t full

[0047] Where a is the rate of decrease of power, and the charging power decreases at a constant rate.

[0048] (1) Analysis of the power change of the electric vehicle during the discharge process:

[0049] The change in the discharge power of an electric vehicle is usually not linear. Especially during the discharge process of the battery, the power changes with the state of the battery. Specifically:

[0050] At the beginning stage of discharge, the power gradually increases from 0 to the maximum power P maxrel , and this stage is represented by the following formula:

[0051]

[0052] Where P′(t) is the discharge power at time t, and P maxrel is the maximum discharge power of the battery, and t ramp is the time required for the discharge power to increase from 0 to P maxrel .

[0053] During the discharge process stage, the battery discharges at the maximum discharge power P maxrel , and this stage is represented by the following formula:

[0054] P′(t) = P maxrel , t ramp <t<t end -t decay

[0055] Where t end is the discharge end time, and t decay is the time required for the discharge power to decrease from P maxrel to 0.

[0056] At the end stage of discharge, the power gradually decreases from the maximum discharge power P maxrel to 0, and this stage can be represented by the following formula:

[0057]

[0058] Step 2, obtain the historical charge and discharge data of each electric vehicle and the corresponding charging pile.

[0059] In this embodiment, the charge and discharge time data and charge and discharge power data of the electric vehicle are obtained from two key data sources, namely the electric vehicle middle platform and the power grid power middle platform, to form the historical charge and discharge data.

[0060] In the present invention, the historical charge and discharge data can be preprocessed, such as data cleaning, conversion, normalization and other operations, to eliminate noise, remove outliers, and fill in missing values, improving the data quality and providing reliable data support for subsequent algorithm processing.

[0061] Step 3: Use the Monte Carlo method to predict the daily total power change curve for a specified date multiple times.

[0062] Refer to Figure 2 As shown, the specific process of this step 3 is as follows:

[0063] Step 301: Set calculation parameters, including the total number of loops M, the total number of electric vehicles N, the number of electric vehicles in each region, the charging pile type, and other parameter values;

[0064] Step 302: Adopt the Monte Carlo method to randomly extract the daily charge and discharge power data of electric vehicles for each region from the historical charge and discharge data set, including the charging start time, charging end time, discharging start time, discharging end time, maximum charging power, minimum charging power, maximum discharging power, and minimum discharging power, and use the power change model of electric vehicles during the charge and discharge process to obtain their power change curves during charging and discharging;

[0065] Step 303: Determine whether the number of extracted electric vehicles is greater than the preset total number of electric vehicles. If so, superimpose and calculate the charge and discharge power change curves of N electric vehicles to obtain the daily total power change curve, and execute Step 304; otherwise, return to Step 302;

[0066] Step 304: Determine whether the current number of loops is greater than the total number of loops. If so, execute Step 305; otherwise, return to Step 302;

[0067] Step 305: Take out the charge power change curves of electric vehicles in each region, calculate the average charging power of the corresponding region, calculate the proportion of the average charging power of each region in the total average charging power of all regions, that is, the charging power proportion of each region is shown in Equation (1); use the number of charging piles in each region as an index to characterize the charging demand, calculate the proportion of the charging demand of each region in the total demand for each region respectively, as the charging demand proportion of each region is shown in Equation (2); after calculating the supply-demand difference component using the charging power proportion and charging demand proportion of the same region, accumulate and add them to obtain the supply-demand difference, and determine whether the supply-demand difference is less than the preset threshold and the supply-demand difference is gradually decreasing. If so, end; otherwise, repeat Step 302.

[0068]

[0069] In the formula, is the average charging power, \(i\) represents the current area, \(N\) is the total number of areas, \(j\) is the area index, \(U\) i represents the proportion of charging demand in the \(i\)-th area, \(Q\) i represents the number of charging piles in the \(i\)-th area.

[0070] In step 302, by extracting the state data of electric vehicles from historical data, the daily charging state time period and daily discharging state time period of electric vehicles are obtained, including the charging start time, charging end time, discharging start time, discharging end time, and combined with the power data when the electric vehicle is connected to the charging pile, the power change data during charging and discharging is obtained, including the maximum charging power, minimum charging power, maximum discharging power, and minimum discharging power.

[0071] In step 303, in each loop, the charging power change curves and discharging power change curves of all electric vehicles are superimposed to obtain the daily total power change curve.

[0072] In step 305, for the spatial distribution stability analysis of charging demand, the way of dividing the spatial area needs to be clarified first. Usually, the spatial area can be divided according to geographical ranges such as communities, streets, and administrative regions. Each area can be abstracted as a node, and the charging demand distribution of the node can be represented by a matrix.

[0073] In step 3 of the present invention, in order to accurately simulate the impact of the charging and discharging behavior of electric vehicles on the power system, all the historical charging and discharging data of electric vehicles in the past five years are selected; according to the number of electric vehicles recorded in these data, the value of \(N\) is set to reflect the scale of the actual electric vehicle group.

[0074] In step 3 of the present invention, the Monte Carlo method is used to randomly extract electric vehicles and their charging and discharging dates, so as to simulate the possible charging and discharging activities of these vehicles on the specified date in each simulation, thereby improving the accuracy of the Monte Carlo simulation; during the \(M\) simulations, the charging and discharging power data of each simulation are recorded in detail for in-depth data analysis and result interpretation. By gradually increasing the number of simulations and continuously observing the stability of the simulation results until the convergence state is reached, the reliability and convergence of the Monte Carlo simulation are ensured.

[0075] In step 3 of the present invention, by setting the proportion of charging demand and comparing the simulated charging power proportion with this charging demand proportion, the simulation tends to the real situation, improving the reliability and robustness of the simulation. Since there is an association between charging and discharging in the actual application of electric vehicles, the present invention only uses the charging power proportion as the determination parameter for the convergence state, reducing the calculation amount while still ensuring the reliability of the simulation.

[0076] Step 4: Analyze the multiple daily total power change curves using the K-means clustering algorithm to obtain the daily total power change curve with the highest probability.

[0077] The K-means algorithm is a typical prototype-based objective function clustering method. It uses a certain distance from the data points to the prototype as the optimization objective function and obtains the adjustment rule of the iterative operation by using the method of finding the extreme value of the function. The K-means algorithm uses the Euclidean distance as the similarity measure. It is to find the optimal classification corresponding to a certain initial clustering center vector V to minimize the evaluation index J. The algorithm uses the sum of squared errors criterion function as the clustering criterion function. Since the value of the clustering model k is unknown, the elbow method is introduced to calculate the most suitable clustering k value.

[0078] The specific principle of the elbow method is as follows: As the number of clusters k increases, the division of the data set will be more detailed, and the similarity of the samples within each cluster will gradually increase, resulting in a continuous decrease in the sum of squared errors (SSE). When the value of k has not reached the optimal number of clusters, increasing k will significantly improve the compactness within the cluster, so the decreasing trend of SSE is significant; however, once the value of k reaches the ideal number of clusters, the improvement effect of the compactness within the cluster brought by further increasing k will quickly weaken, which is reflected in the sharp slowdown of the decreasing rate of SSE. Subsequently, as the value of k continues to increase, the downward path of SSE will tend to be stable. In other words, the relationship graph between SSE and k presents an elbow shape, and the inflection point of this shape, that is, the position of the elbow, indicates the optimal number of clusters of the data set.

[0079] In step 4 of the present invention, for the multiple daily total power change curves of electric vehicles obtained by the Monte Carlo simulation method, vectorization processing is first implemented to meet the requirements of mathematical operations and pattern recognition; then a clustering algorithm is used to divide these multiple vectorized daily total power change curves of electric vehicles. Based on the convergence of the Monte Carlo simulation algorithm, a dominant category containing the vast majority of curves and several minor categories containing fewer curves can be successfully identified.

[0080] Based on the clustering simulation calculation results, select the dominant cluster group in the clustering results. Use the cluster center of this dominant cluster group as the vector representation of the ideal daily total power curve, and convert the vector representation of the cluster center to obtain a comprehensive electric vehicle charging and discharging power change curve accordingly. This curve depicts the charging and discharging habits of most electric vehicles on a specified date and provides crucial reference information for the dispatching and planning of the power system.

[0081] In the existing electric vehicle charging and discharging power prediction methods, regression calculations are mostly carried out by collecting existing data. However, due to the randomness of electric vehicle charging and discharging behaviors, the prediction accuracy of this charging and discharging power prediction method is not high. In the present invention, the Monte Carlo algorithm is set to simulate the daily charging and discharging power of electric vehicles. When the number of simulations is large enough, its accuracy tends to approach the actual occurrence. On this basis, the present invention combines the clustering algorithm to evaluate the future daily total power change of electric vehicles, so as to select the most likely daily total power change situation of electric vehicles, effectively improving the prediction level of electric vehicle charging and discharging power, and greatly enhancing the photovoltaic power consumption and the power supply capacity of the power grid during peak periods.

[0082] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm, characterized in that: The method comprises: constructing a power variation model of the electric vehicle during the charging and discharging process; Obtain historical charging and discharging data of each electric vehicle and the corresponding charging pile, including power change calculation parameters; The Monte Carlo method is used to predict the daily total power change curve of the specified date multiple times, wherein the daily total power change curve is obtained by calculating and summarizing the historical charging and discharging data of randomly selected electric vehicles according to the power change model; A clustering algorithm is used to analyze the multiple daily total power change curves to obtain the daily total power change curve with the highest probability.

2. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 1, characterized in that: The charging power calculation model of the electric vehicle is expressed as: Where P(t) represents the charging power at time t; P max represents the peak value of charging power; k is the parameter that determines the power growth rate; t max Indicates the time required for charging power to reach peak value, t fall Indicates the charging power from t max The time required to drop to 0, t full Indicates the charging end time.

3. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 1, characterized in that: The discharge power calculation model of the electric vehicle is expressed as: Where P maxrel is the maximum discharge power of the electric vehicle battery, t ramp As the discharge power increases from 0 to P maxrel The time required, t decay The discharge power is from P maxrel The time required to decrease to 0, t end The discharge end time.

4. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in any one of claims 1 to 3, It is characterized in that The method of using the Monte Carlo method to predict the charging and discharging power of the electric vehicle on a specified date multiple times includes: Initial setting step, setting calculation parameters, including the total number of cycles M and the total number of electric vehicles N; A random extraction step, randomly extracting the status data of N electric vehicles from historical data, obtaining the charging and discharging status time period of the electric vehicles in one day, combining the data when the electric vehicles are connected to the charging pile, and using the power change model to obtain the power change curve of each electric vehicle during charging and discharging, superimposing and calculating the charging and discharging power change curves of all electric vehicles, and obtaining the daily total power change curve; The loop step repeats the random sampling step M times in total to obtain M daily total power change curves.

5. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 4, characterized in that: The Monte Carlo method predicts the charging and discharging power of electric vehicles on a specified date multiple times, and also includes a convergence determination step: after obtaining M daily total power change curves, the electric vehicle charging power change curves in each area are taken out, the charging power proportion of each area is calculated, and the supply and demand difference is analyzed using the charging power proportion and the charging demand proportion of each area. When the supply and demand difference is less than a preset threshold and shows a decreasing trend, convergence is determined, otherwise the random extraction step and the loop step are repeated.

6. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 5, characterized in that: The supply-demand difference is obtained by accumulating the difference between the charging power ratio and the charging demand ratio in each area.

7. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 6, characterized in that: The proportion of charging power in each area is expressed as: In the formula, is the average charging power, i represents the current area, N is the total number of areas, and j is the area index.

8. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 6, characterized in that: In the above-mentioned proportion of charging demand, the number of charging piles in each area is used as an indicator to characterize the charging demand. For each area, the proportion of its charging demand to the total demand is calculated and expressed as: Among them, U i represents the proportion of charging demand in the ith area, Q i represents the number of charging piles in the ith area, and N is the total number of areas.

9. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 1, characterized in that: In the clustering algorithm, all daily charge and discharge total power change curves are classified, the dominant category of the curve is determined according to the majority principle, and the daily charge and discharge total power change curve under the dominant category is used for data fitting to obtain the ideal daily charge and discharge power change curve of the electric vehicle.

10. A method for quantitatively evaluating the charging and discharging state of an electric vehicle based on a simulation algorithm as claimed in claim 9, characterized in that: In the data fitting, the central point of the dominant category is converted to serve as the daily total power variation curve of the electric vehicle with the highest probability.

Images