With the rapid proliferation of electric vehicles in China, the charging demand poses significant challenges to the stability of power systems. Uncontrolled charging of electric vehicles can lead to overutilization of charging infrastructure, causing sharp fluctuations in grid load during peak hours and threatening grid security. In this study, we explore a power allocation-based orderly charging strategy to address the chaotic management of charging facilities. By considering user charging needs and load fluctuations, we adjust charging power across different time slots to achieve orderly management. This approach not only reduces the peak-valley difference in grid load but also lowers operational costs and enhances grid efficiency. Furthermore, it disperses charging loads, improves economic benefits, and supports the sustainable development of electric vehicles in China.
The growing adoption of China EV has intensified the need for efficient charging solutions. As more consumers switch to electric vehicles, the strain on electrical grids during high-demand periods becomes apparent. This paper investigates key factors in charging decision-making, analyzes the diversity of charging requirements, and proposes an orderly charging regulation strategy under adjustable power. Through an adaptive angular region division method, we optimize the charging process, thereby increasing charging efficiency and grid stability. We construct an orderly charging model with defined objective functions and constraints, and conduct case studies to validate the model’s effectiveness. Our research demonstrates that rational charging strategies can meet user demands while promoting the sustainable growth of electric vehicles.
Electric vehicle charging decisions are influenced by various factors, including user travel habits, vehicle usage frequency, availability of charging infrastructure, and battery capacity. To quantify the charging demand for electric vehicles, we use the following formula: $$ Q = C \times D $$ where \( Q \) represents the total charging demand, i.e., the required charging energy within a specific time frame; \( C \) denotes the battery capacity, reflecting the total electrical energy the electric vehicle can store; and \( D \) is the average daily driving distance of the electric vehicle. By establishing a charging demand model, we can deeply analyze the characteristics of electric vehicle charging behavior, providing a foundation for orderly charging strategies. The diversity in charging demand necessitates tailored approaches, as users in different scenarios, such as residential areas or commercial zones, have varying priorities.
Charging methods for electric vehicles primarily include slow charging and fast charging. Slow charging typically occurs during nighttime or periods of low electricity demand, utilizing off-peak tariffs to charge electric vehicles. This method causes less battery degradation, helping to extend battery lifespan. Users can charge their vehicles overnight and have a full battery by morning, avoiding daytime charging queues. However, slow charging has the drawback of taking several hours to fully charge the battery, making it suitable for daily commutes or non-urgent scenarios. Fast charging, on the other hand, is used during daytime or when rapid energy replenishment is needed. Fast charging equipment can deliver substantial energy to electric vehicles in a short time, such as charging a certain percentage of the battery within 30 minutes to 1 hour. This allows users to charge quickly at specific locations, enhancing vehicle flexibility. The disadvantages of fast charging include higher costs compared to slow charging and potential battery wear from frequent use, which could impact long-term battery health. To manage electric vehicle charging demand and optimize charging methods, we employ an orderly charging strategy based on time-of-use tariffs, which guides users to charge during low-tariff periods and effectively reduces grid load, promoting rational utilization of power resources. The dynamic pricing model is expressed as: $$ P_c = P_b (1 + \alpha E_p – \beta E_v) $$ where \( P_c \) is the electricity price during charging periods; \( P_b \) is the base price without peak-valley adjustments; \( \alpha \) and \( \beta \) are adjustment coefficients for peak and off-peak prices, respectively, reflecting changes in electricity demand across different time slots; and \( E_p \) and \( E_v \) are proportion factors for peak and off-peak prices, quantifying the differences in tariffs.
In residential areas, where users are predominantly commuters, charging times are concentrated from 18:00 after work until 08:30 the next morning. These users prioritize charging costs and battery life over charging speed. Therefore, residential charging strategies must account for peak-valley tariff differences to optimize charging expenses. In commercial areas and highway service zones, electric vehicle users primarily require fast charging to quickly replenish energy in a short time. Users in these regions focus more on charging rates, so a “first-come, first-served” orderly charging strategy is adopted to ensure charging efficiency. To reduce the total charging load within segmented time periods and achieve orderly charging, we set a maximum limit load that the grid can withstand: the sum of residential electricity load and charging load. By controlling the number of charging vehicles, we prevent excessive concentration of grid load. Each day is divided into 96 time slices of 15 minutes each. Every 15 minutes, the current total charging load of vehicles is calculated. If the load exceeds the limit, charging power is regulated to reduce the load and ensure stable grid operation.
In optimizing the orderly charging strategy for electric vehicles, we address the issue of traditional Multi-Objective Particle Swarm Optimization (MOPSO) algorithms overemphasizing solution feasibility, leading to over-exploitation of feasible regions and neglect of information from infeasible regions, which affects optimization outcomes. We partition the target space into multiple regions based on angles, with each region corresponding to a specific angle, thereby altering the search direction of particles in the target space. This method enables particles to follow more complex search paths and discover optimal solutions more effectively. As the number of particles increases, their angle values become closer, resulting in more intricate search paths and enhanced diversity. The particle update process is described by the following formulas. The velocity update formula is: $$ v_i = \omega v_i + c_1 r_1 (p_{\text{best}} – p_i) + c_2 r_2 (g_{\text{best}} – p_i) $$ where \( v_i \) is the velocity of particle \( i \); \( \omega \) is the inertia weight; \( c_1 \) and \( c_2 \) are learning factors; \( r_1 \) and \( r_2 \) are random numbers between [0, 1]; \( p_i \) is the position of particle \( i \) in iteration \( r \); \( p_{\text{best}} \) is the individual best position; and \( g_{\text{best}} \) is the global best position. The position update formula is: $$ p_i = p_i + v_i $$ Before dividing the angular regions, a reference point must be determined. The origin (0,0) is selected as this reference point. After establishing the reference point, initial angle division is performed, partitioning the target space around the origin into three equal angular regions, each representing a specific angle range in the target space. As division iterations increase, the size of each angular region is gradually reduced. This process involves subdividing each initial region into smaller subregions iteratively until the desired precision level is achieved. After the angular regions are subdivided, an intensive search is conducted for each small region on the target axis.
When constructing the orderly charging model for electric vehicles under adjustable power, the objective function defines the optimization direction. The goal of minimizing charging costs (\( f_1 \)) aims to reduce the total charging costs for all electric vehicle users. The calculation of charging costs considers electricity prices and charging power across different charging periods, with the specific formula as follows: $$ f_1 = \sum_{i} \sum_{j} P_{ij} t_{ij} C_j $$ where \( i \) is the index of the electric vehicle user; \( j \) is the index of the charging time slot; \( P_{ij} \) is the charging power of user \( i \) in time slot \( j \); \( t_{ij} \) is the charging time of user \( i \) in time slot \( j \); and \( C_j \) is the electricity price in time slot \( j \). This objective function reduces overall electricity expenses for users by reasonably scheduling charging times and power allocation. The goal of minimizing peak load (\( f_2 \)) aims to lower the grid’s peak load, improve overall utilization efficiency, and ensure power system stability. The peak load calculation formula is: $$ f_2 = \max_t \left( L_t + \sum_{i} P_{ij} \right) $$ where \( L_t \) is the base load of the grid at time \( t \), and \( P_{ij} \) is the charging power of user \( i \) at time \( t \). By adjusting charging power and time, the grid load during peak periods can be effectively reduced, alleviating grid pressure and enhancing its capacity.
Constraints are essential to ensure the feasibility and safety of the charging model. The total charging power of each charging pile must not exceed its maximum capacity, preventing overload damage and ensuring equipment safety. If \( P_{ij} \) represents the charging power of user \( i \) at charging pile \( j \), the constraint is: $$ \sum_{i=1}^{n} P_{ij}^t \leq P_j^{\text{max}} \quad \forall j, t $$ where \( P_{ij}^t \) is the charging power of user \( i \) at charging pile \( j \) at time \( t \), and \( P_j^{\text{max}} \) is the maximum capacity of charging pile \( j \). The total grid load must be less than its maximum load limit to ensure safe and stable grid operation. This constraint is expressed as: $$ P_s^t = L_t + \sum_{i=1}^{n} P_{\text{ev},i}^t \leq P^{\text{lim}} \quad \forall t $$ where \( L_t \) is the residential load at time \( t \), \( n \) is the number of electric vehicles, \( P_{\text{ev},i}^t \) is the charging power of the \( i \)-th electric vehicle at time \( t \), and \( P^{\text{lim}} \) is the maximum capacity of the distribution network.
The orderly charging regulation strategy involves a structured process. When an electric vehicle connects to a charging station, the charging pile obtains key vehicle information through onboard communication systems. Based on the acquired vehicle data and historical charging records, charging demand is predicted using statistical and machine learning methods. An improved multi-objective particle swarm optimization algorithm is employed to design the power allocation scheme for charging piles. According to the power allocation plan, the output power of charging piles is adjusted in real-time, while charging status and grid load are monitored. After charging is complete, the power output of the charging pile is automatically stopped, and the charging end time is recorded. The charging process is documented in detail, and charging effectiveness is evaluated to provide data for charging regulation. Through these steps, orderly charging modeling ensures efficient operation of electric vehicle charging stations, minimizes impact on the grid, and enhances charging safety. This modeling approach not only improves the operational efficiency of charging stations but also contributes to grid stability and user experience.
For instance analysis, we consider a residential community with 150 households, where 40% of residents own electric vehicles, resulting in 60 electric vehicles. To build an effective charging model, we set the parameters as shown in Table 1.
| Parameter | Value |
|---|---|
| Residential Load Limit (kW) | 1000 |
| Number of Electric Vehicles | 60 |
| Number of Charging Piles | 60 |
| Rated Power of Charging Pile (kW) | 7 |
| Slow Charging Time for Electric Vehicles (h) | 8 |
Assuming the distribution transformer in the residential area has a rated power of 1000 kW, and in load simulation, the base load is set to \( y \), we use the Monte Carlo method to simulate the total load and base load of the residential area. Results show that the total load can reach approximately 1200 kW during peak periods, while the base load is about 600 kW, revealing the high-demand characteristics of electricity in specific time slots. Based on the simulation, electricity peaks in the residential area mainly occur from 18:00 to 21:00, when household electricity demand rises, causing overall power demand to surge. By simulating uncontrolled charging of electric vehicles, we find that charging peaks for electric vehicles also overlap from 17:00 to 20:00, coinciding with household electricity peaks. This overlap increases load pressure on the distribution network, affecting power supply stability.
To address this issue, we propose an orderly charging regulation strategy. The core of this strategy is to reasonably allocate charging time and power through real-time monitoring and scheduling, ensuring that charging demands are met without exceeding the distribution transformer’s rated power. Based on an analysis of electric vehicle owners’ charging habits, we divide the charging time for 60 electric vehicles into three periods: peak period (17:00–20:00), normal period (20:00–23:00), and off-peak period (23:00–07:00). During the peak period, each vehicle’s charging power is limited to 3 kW to ensure the total load does not exceed the transformer’s safe rating. In the normal period, charging power increases to 5 kW, and during the off-peak period, full charging power of up to 7 kW is allowed. We introduce smart meters and real-time monitoring systems to track electricity demand and charging status in the community. When the total load approaches 1000 kW, the system automatically adjusts the charging power or time of electric vehicles to prevent overload. Additionally, the system regularly sends charging status notifications and suggestions to residents, encouraging them to charge during off-peak periods.
The effectiveness of the orderly charging regulation strategy is evaluated over one month of monitoring. The assessment results are as follows:
- Load Balancing: Before implementing orderly charging, the community’s peak load reached 1200 kW. After implementation, the peak load was successfully controlled within 950 kW, effectively alleviating power supply pressure.
- User Satisfaction: Through surveys, 92% of residents expressed satisfaction with the charging regulation measures, stating that it ensured charging convenience while reducing congestion during peak electricity hours and improving overall living comfort.
- Charging Efficiency: Prior to implementing the orderly charging strategy, the charging efficiency of electric vehicles was about 70%. After implementation, charging efficiency increased to 85%, as reasonable time allocation and power management reduced charging delays caused by insufficient power.
- Economic Benefits: Overall economic benefits improved significantly. By limiting charging power during peak periods, the community’s electricity expenses decreased by approximately 15%, and maintenance costs due to overload were reduced.
These results highlight the advantages of the orderly charging strategy for electric vehicles in China, demonstrating its potential to enhance grid stability and user experience while supporting the growth of the China EV market.
In conclusion, as electric vehicles become more widespread, the demand for charging infrastructure continues to grow. This paper thoroughly analyzes charging demand, charging methods, and various regulation strategies for power allocation. Charging demand considers factors such as user behavior, charging time, and charging costs. In terms of power allocation, we propose an orderly charging strategy under adjustable power, providing optimized solutions for electric vehicle charging. This strategy not only improves charging efficiency but also reduces impact on the grid, ensuring a balance between charging demand and power supply. In constructing the orderly charging model, we define objective functions and constraints to make the charging process more controllable. The introduction of the adaptive angular region division method offers insights for charging scheduling, enabling precise responses to charging demands in different areas. Through case analysis, we validate the effectiveness of the proposed strategy, providing support for the stable operation of power systems and the sustainable development of electric vehicles in China.