A system dynamics-based electric vehicle charging load prediction method and system

Abstract

The application discloses a kind of electric vehicle charging load prediction method and system based on system dynamics, belong to electric vehicle load prediction technical field.Method includes as follows: first, load historical data, with the aid of Gompertz function, exponential decay function predicts vehicle population, generates carbon emission factor and pre-processes sequence;Again, through oil-electricity car life cycle cost, market utility dynamic simulation fleet inventory evolution;Then, in combination with driving, energy consumption parameter calculates driving charging, V2G discharge and replenishment charging load;Finally, build baseline, active V2G two types of scene simulation, visual output comparison curve is completed analysis.System is equipped with four major functional modules, Python is realized, built-in data missing fault-tolerant mechanism.The application constructs closed-loop feedback model, solves the problem that traditional model prediction deviation is big, scene adaptation is poor, can provide decision support for power grid planning.

Description

Technical Field

[0001] This invention relates to the fields of electric vehicle big data analysis, system dynamics modeling, and electric vehicle charging load prediction technology, and particularly to an electric vehicle charging load prediction method and system based on system dynamics. Background Technology

[0002] With the rapid popularization of electric vehicles, their charging behavior exhibits characteristics of large scale and high spatiotemporal uncertainty, posing a significant challenge to the safe and stable operation of the power grid. Traditional electric vehicle charging load forecasting methods mainly rely on statistical analysis or simple machine learning models, which can only capture the correlation of historical data. They cannot dynamically simulate the long-term impact of policy interventions (such as carbon trading, subsidy continuation, orderly charging, and vehicle-to-grid technology) on the size of the electric vehicle fleet and charging load. They also struggle to handle the competitive relationship between electric vehicles and gasoline vehicles and the cost reduction brought about by technological advancements. Therefore, when forecasting the medium- and long-term electric vehicle charging load under the influence of high-dimensional mixed variables, the lack of a dynamic feedback mechanism leads to technical problems such as large prediction bias, poor scenario adaptability, and weak decision support capabilities.

[0003] Therefore, it is necessary to improve the traditional electric vehicle charging load prediction method. Summary of the Invention

[0004] To address the technical problems of existing technologies in predicting medium- and long-term electric vehicle charging load under the influence of high-dimensional mixed variables, which result in large prediction deviations, poor scenario adaptability, and weak decision support capabilities due to the lack of dynamic feedback mechanisms, the present invention aims to propose an electric vehicle charging load prediction method and system based on system dynamics.

[0005] To achieve the objectives of this invention, the technical solution provided by this invention is as follows: First aspect This invention provides a method for predicting electric vehicle charging load based on system dynamics, comprising the following steps: Step 1: Load historical data, predict vehicle ownership, generate orderly charging, vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing; if there are missing data files, output a warning and call placeholder data to run, and output a well-organized preprocessed data matrix. Step 2: Receive the preprocessed data matrix, and based on lifecycle cost calculation, market share prediction and policy parameters, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory, and output fleet size, lifecycle cost and electric vehicle market share results data; Step 3: Receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharge load and supplementary charging load, and output a multi-scenario load data table; Step 4: Run the baseline scenario and the active vehicle to grid scenario respectively, plot the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

[0006] Furthermore, in step 1, the Gompertz function is used to calculate the proportion of vehicles, and the total number of vehicles is calculated in combination with population data.

[0007] Furthermore, in step 1, an exponential decay function is used to fit the carbon emission factor of the power grid, and interpolation is used to generate a complete carbon emission factor sequence from 2010 to 2060.

[0008] Furthermore, in step 2, the purchase cost, resale value, maintenance cost, usage cost, and carbon trading cost of electric vehicles and fuel vehicles are calculated annually. Additionally, the battery replacement cost of electric vehicles, orderly charging, and net revenue from vehicle grid connection are calculated to obtain the complete life cycle cost of the two types of vehicles.

[0009] Furthermore, in step 2, the market utility of electric vehicles is calculated using market sensitivity parameters, license plate restrictions, and environmental factors; historical sales share is used before 2022, and annual sales of hybrid vehicles are updated based on market utility after 2023, and the total vehicle inventory is updated in conjunction with vehicle scrapping volume.

[0010] Furthermore, in step 3, the annual charging energy and total driving charging load of a single vehicle are first calculated, and then the supplementary charging load corresponding to the vehicle's discharge to the grid is calculated based on the charging and discharging efficiency.

[0011] Furthermore, step 4 sets up two types of simulation scenarios: the baseline scenario has no vehicle-to-grid revenue and no carbon trading mechanism, and the vehicle purchase subsidy will end after 2023; the active vehicle-to-grid scenario has a participation rate of 50% and an average annual discharge of 1,500 kWh per vehicle.

[0012] Furthermore, this method is implemented in Python, supports automatic loading of CSV data, has a built-in data missing exception handling mechanism, and the visualization curve supports Chinese character display.

[0013] Second aspect This invention provides an electric vehicle charging load prediction system based on system dynamics, used to execute the electric vehicle charging load prediction method based on system dynamics, including a data preparation and preprocessing module, an electric vehicle fleet prediction module, a charging and discharging load calculation module, and a result visualization and output module. The data preparation and preprocessing module is used to load historical data, predict vehicle ownership, generate orderly charging and vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing; if there are missing data files, it outputs a warning and calls placeholder data to run, and outputs a well-organized preprocessed data matrix. The electric vehicle fleet prediction module is used to receive a preprocessed data matrix, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory based on life cycle cost calculation, market share prediction and policy parameters, and output fleet size, life cycle cost and electric vehicle market share results data. The charging and discharging load calculation module is used to receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharging load and supplementary charging load, and output a multi-scenario load data table. The results visualization and output module is used to run the baseline scenario and the active vehicle to grid scenario respectively, draw the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: The method provided in this invention predicts vehicle ownership and adoption rate using the Gompertz model, combines this with lifecycle cost calculation, constructs a dynamic evolution model of electric vehicle and gasoline vehicle fleets, and then overlays charging and discharging load calculation modules to achieve accurate prediction of charging load under multiple policy scenarios. The system is implemented in Python and supports automatic loading of CSV data, anomaly handling, and visualization output. Specifically: First, this invention adopts a system dynamics framework to capture the closed-loop feedback between electric vehicle ownership, sales, scrapping rate and load, enabling long-term dynamic simulation from 2010 to 2060, far exceeding traditional static models. Secondly, it has built-in switches for parameters such as carbon trading, subsidy continuation, orderly charging, and vehicle-to-grid adoption rate, and supports comparisons between baseline and active vehicle-to-grid scenarios to quantify the impact of policy heterogeneity.

[0015] In addition, the built-in placeholder mechanism and exception handling for missing files ensure that the model can still run even when the actual data is incomplete. Attached Figure Description

[0016] Figure 1 A schematic diagram of the system dynamics prediction method for electric vehicle charging load provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that the acquisition of data and collection of information in this application are legal, compliant, or obtained with the consent of the subject of the data collection.

[0019] like Figure 1 As shown, this invention proposes a system dynamics prediction method for electric vehicle charging load, which includes the following steps: Step 1: Load historical data, predict vehicle ownership, generate orderly charging, vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing; if there are missing data files, output a warning and call placeholder data to run, and output a well-organized preprocessed data matrix. Specifically as follows: Step S1: First, define the simulation year range as 2010 to 2060, and determine the year range for the final plot as 2025 to 2060.

[0020] Step S2: Load carbon price forecast data, read the original carbon price sequence from the file, fill in 0 values ​​for 2010-2019, and multiply the subsequent years by 2 to generate the carbon price sequence under the policy scenario.

[0021] Step S3: Load population, GDP, and GDP per capita data, and use the Gompertz function to calculate the annual vehicle ownership percentage. The calculation formula is as follows: ; in, This represents the saturation stock ratio. and For model parameters, For the first Annual GDP per capita.

[0022] Then, The total number of vehicles is calculated by multiplying the population data of that year by the following formula: ; in, For the first Annual population (100 million people) The unit is 100 million vehicles.

[0023] Step S4: Load historical data on total vehicle sales, hybrid vehicle ownership, and vehicle survival rate. For the period from 2010 to 2022, directly use historical ownership and sales data. For years after 2023, calculate the total scrapping volume and total sales volume cyclically based on the previous year's total ownership, current total sales volume, and scrapping rate distributed by vehicle age (maximum consideration of 20 years) to ensure that the total ownership remains balanced.

[0024] Step S5: Fit the grid carbon emission factor using an exponential decay function. The calculation formula is as follows: ; in, , , For fitting parameters, For a point in time.

[0025] After obtaining the parameters through curve fitting, interpolation is performed on the data from 2010 to 2060 to generate a complete sequence (the values ​​before 2010 are fixed as the 2020 values, and the values ​​after 2060 are fixed as the 2060 values).

[0026] Step S6: Generate the S-curves for the adoption rates of ordered charging (OC) and vehicle-to-grid (V2G) technologies, calculated using the following formula: ; in, To standardize the time variables, OC will start in 2025 and V2G will start in 2030, with the adoption rate gradually increasing from 0 to close to 1.

[0027] Step S7: Interpolate the historical sales share data of electric vehicles to generate the target electric vehicle proportion sequence (25% in 2025, 40% in 2030, 60% in 2035, and 90% in 2060).

[0028] Step S8: If any data file is missing, the system will automatically output a warning message and continue running using preset placeholder data to ensure that the entire process is not affected. Finally, it will output a well-organized preprocessed result for use by subsequent modules.

[0029] Step 2: Receive the preprocessed data matrix, and based on lifecycle cost calculation, market share prediction and policy parameters, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory, and output fleet size, lifecycle cost and electric vehicle market share results data; Specifically as follows: Step S21: Input the defined list of years, various policy parameters (whether carbon trading is active and its start year, carbon price series, annual revenue from orderly charging, annual revenue from vehicles connected to the grid, whether subsidies are continued, total revenue from battery recycling, etc.), global technical parameters (electricity price, oil price, vehicle lifespan, depreciation rate, discount rate, carbon quota reduction rate, gasoline carbon conversion factor, etc.), and preprocessed data.

[0030] Step S22: Initialize the inventory status of electric vehicles and gasoline vehicles, starting from the historical inventory in 2010; construct a research and development sub-model, with the initial technology level set at 0.5 and the innovation capability starting from 0. Adjust government R&D investment and enterprise R&D intensity annually based on the gap between the electric vehicle sales share and the target ratio, update the technology level, and calculate the reduction coefficient of electric vehicle production cost accordingly (the higher the technology level, the lower the cost coefficient).

[0031] S23: Calculate the life cycle cost (LCC) of electric vehicles and gasoline vehicles in detail year by year. This application innovatively designs a calculation method, specifically including: Purchase cost: ; in, For production costs, For profit margin, As a subsidy; Resale value: ; in, For depreciation rate, For vehicle lifespan, For purchase costs, The discount rate; Maintenance costs (EV): ; in, Average annual mileage For the current year, For EV lifespan; Maintenance Costs (ICEV): ; in, For ICEV lifespan; Battery replacement cost (EV): ; Usage cost: ; in, Average annual mileage Energy consumption rate For electricity or oil prices; Carbon trading costs (only when carbon trading is active): ; in, Annual carbon emissions, For carbon quotas, For carbon price, As a promoting factor; Net revenue from orderly charging and vehicle-to-grid connections.

[0032] Ultimately forming a complete life cycle cost and .

[0033] S24: This application innovatively designs a method for calculating the market utility and market share of electric vehicles based on market sensitivity parameters (electric vehicle sensitivity, fuel vehicle sensitivity, environmental factors, license plate restriction impact, etc.). The utility calculation formula is as follows: ; in, , The sensitivity coefficient, , For life cycle cost, Due to the impact of license restrictions, , Environmental factors. Historical sales share data will be mandatory until 2022; thereafter, annual sales of electric vehicles and gasoline vehicles will be dynamically calculated and updated based on LCC differences.

[0034] S25: Calculate the annual scrap volume based on vehicle age distribution and scrap rate, continuously update the total inventory of electric vehicles and gasoline vehicles, and finally output a complete result data table including the total inventory of electric vehicles, the total inventory of gasoline vehicles, the life cycle cost of electric vehicles, the life cycle cost of gasoline vehicles, and the market share of electric vehicles.

[0035] Step 3: Receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharge load and supplementary charging load, and output a multi-scenario load data table; Step S31: Input the fleet prediction results data, the annual average mileage sequence (which starts from 10,000 km and increases linearly with the year), the energy consumption rate sequence (electricity consumption decreases logarithmically with the year, and fuel consumption decreases power law with the year), the charging efficiency factor (default 0.90), the vehicle-to-grid discharge efficiency factor (default 0.85), and the vehicle-to-grid and orderly charging adoption rate sequence.

[0036] Step S32: First, calculate the baseline driving-related charging energy for each electric vehicle. The calculation formula is as follows: ; in, Energy consumption rate, For charging efficiency.

[0037] Step S33: Calculate the total driving charging load, the calculation formula is as follows: ; in, Total inventory of electric vehicles (millions of vehicles). The orderly charging influence factor is converted to gigawatt-hours (GWh).

[0038] Step S34: For the vehicle-to-grid (V2G) portion, an innovative approach was adopted: the number of vehicles with V2G capability was calculated based on the adoption rate, then multiplied by the active participation ratio and the average annual discharge volume per active vehicle to obtain the total V2G discharge load; then, the additional charging load required for supplementary charging was calculated using the following formula: ; in, For vehicles to discharge loads to the power grid, For discharge efficiency, For charging efficiency.

[0039] S35: The final output includes a result data table containing total charging load, vehicle-to-grid discharge load, driving charging load, and vehicle-to-grid supplementary charging load. It supports baseline scenarios (vehicle-to-grid participation rate of 0) and active vehicle-to-grid scenarios (participation rate of 50%, annual average discharge of 1500 kWh).

[0040] Step 4: Run the baseline scenario and the active vehicle to grid scenario respectively, plot the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

[0041] Step S41: Run the baseline scenarios (no vehicle-to-grid revenue, no carbon trading, and subsidies ending after 2023) separately to generate baseline load results.

[0042] Step S42: Run active vehicles to the grid scenario (participation rate 50%, annual average discharge 1500 kWh) and generate active scenario load results.

[0043] Step S43: Use the plotting tool to plot the curve of electric vehicle inventory changing over time, as well as the comparison curves of total charging load and vehicle discharge load to grid under the two scenarios (Chinese display is supported).

[0044] Step S44: Output sample data for a specific year (e.g., 2030) to the console, including electric vehicle inventory, life cycle cost, market share, and load values. Then, by adjusting policy parameters and global parameters, perform comparative analysis and decision support under different policy scenarios.

[0045] In addition, the present invention provides an electric vehicle charging load prediction system based on system dynamics, for executing the electric vehicle charging load prediction method based on system dynamics, including a data preparation and preprocessing module, an electric vehicle fleet prediction module, a charging and discharging load calculation module, and a result visualization and output module. The data preparation and preprocessing module is used to load historical data, predict vehicle ownership, generate orderly charging and vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing; if there are missing data files, it outputs a warning and calls placeholder data to run, and outputs a well-organized preprocessed data matrix. The electric vehicle fleet prediction module is used to receive a preprocessed data matrix, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory based on life cycle cost calculation, market share prediction and policy parameters, and output fleet size, life cycle cost and electric vehicle market share results data. The charging and discharging load calculation module is used to receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharging load and supplementary charging load, and output a multi-scenario load data table. The results visualization and output module is used to run the baseline scenario and the active vehicle to grid scenario respectively, draw the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

[0046] Finally, it should be noted that the above embodiments are merely illustrative and explanatory of the present invention, and are not intended to limit the present invention to the scope of the described embodiments. Furthermore, those skilled in the art will understand that the present invention is not limited to the above embodiments, and many more variations and modifications can be made based on the teachings of the present invention, all of which fall within the scope of protection claimed by the present invention.

Claims

1. A method for predicting electric vehicle charging load based on system dynamics, characterized in that, Includes the following steps: Step 1: Load historical data, predict vehicle ownership, generate orderly charging, vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing. If a data file is missing, output a warning and run the program using placeholder data, outputting a well-organized preprocessed data matrix. Step 2: Receive the preprocessed data matrix, and based on lifecycle cost calculation, market share prediction and policy parameters, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory, and output fleet size, lifecycle cost and electric vehicle market share results data; Step 3: Receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharge load and supplementary charging load, and output a multi-scenario load data table; Step 4: Run the baseline scenario and the active vehicle to grid scenario respectively, plot the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

2. The electric vehicle charging load prediction method based on system dynamics according to claim 1, characterized in that, In step 1, the Gompertz function is used to calculate the proportion of vehicles, and the total number of vehicles is calculated by combining the population data.

3. The electric vehicle charging load prediction method based on system dynamics according to claim 2, characterized in that, In step 1, the exponential decay function is used to fit the carbon emission factor of the power grid, and interpolation is used to generate a complete carbon emission factor sequence from 2010 to 2060.

4. The electric vehicle charging load prediction method based on system dynamics according to claim 2, characterized in that, Step 2 calculates the purchase cost, resale value, maintenance cost, usage cost, and carbon trading cost of electric vehicles and gasoline vehicles year by year. It also calculates the battery replacement cost of electric vehicles, orderly charging, and net revenue from vehicle grid connection to obtain the complete life cycle cost of the two types of vehicles.

5. The electric vehicle charging load prediction method based on system dynamics according to claim 4, characterized in that, In step 2, the market utility of electric vehicles is calculated using market sensitivity parameters, license plate restrictions, and environmental factors. Before 2022, historical sales share is used, and after 2023, the annual sales volume of hybrid vehicles is updated based on market utility, and the total vehicle inventory is updated in conjunction with the vehicle scrapping volume.

6. The electric vehicle charging load prediction method based on system dynamics according to claim 5, characterized in that, In step 3, the annual charging energy and total driving charging load of a single vehicle are calculated first, and then the supplementary charging load corresponding to the vehicle's discharge to the grid is calculated based on the charging and discharging efficiency.

7. The electric vehicle charging load prediction method based on system dynamics according to claim 6, characterized in that, Step 4 sets up two types of simulation scenarios: the baseline scenario has no vehicle-to-grid revenue and no carbon trading mechanism, and the vehicle purchase subsidy will end after 2023; the active vehicle-to-grid scenario has a participation rate of 50% and an average annual discharge of 1,500 kWh per vehicle.

8. The electric vehicle charging load prediction method based on system dynamics according to claim 7, characterized in that, This method is implemented in Python, supports automatic loading of CSV data, has a built-in data missing exception handling mechanism, and the visualization curve supports Chinese characters.

9. A system for predicting electric vehicle charging load based on system dynamics, characterized in that, The method for performing electric vehicle charging load prediction based on system dynamics as described in any one of claims 1 to 8 includes a data preparation and preprocessing module, an electric vehicle fleet prediction module, a charging and discharging load calculation module, and a result visualization and output module. The data preparation and preprocessing module is used to load historical data, predict vehicle ownership, generate orderly charging, vehicle-to-grid technology adoption rate curves and grid carbon emission factor sequences, and carry out data cleaning and standardization processing. If a data file is missing, output a warning and run the program using placeholder data, outputting a well-organized preprocessed data matrix. The electric vehicle fleet prediction module is used to receive a preprocessed data matrix, dynamically simulate the evolution of electric vehicle and fuel vehicle fleet inventory based on life cycle cost calculation, market share prediction and policy parameters, and output fleet size, life cycle cost and electric vehicle market share results data. The charging and discharging load calculation module is used to receive the fleet prediction results, combine driving mileage, energy consumption rate, and technology adoption rate under different scenarios, calculate driving charging load, vehicle-to-grid discharging load and supplementary charging load, and output a multi-scenario load data table. The results visualization and output module is used to run the baseline scenario and the active vehicle to grid scenario respectively, draw the electric vehicle inventory curve and the multi-scenario load comparison curve, output sample data to the console, and complete the multi-scenario comparative analysis.

Images