MODELING CLUSTER-BASED CONTROLLED LOAD SHARING FOR SCALABLE ELECTRIC VEHICLES

ABSTRACT

In this paper, a cluster model of controlled charging load based on user behavior characteristics is proposed for four typical scenarios: residential areas, commercial areas, parks, and centralized charging stations. By analyzing the network characteristics and the characteristics of the normal load in different scenarios, a unified model of controlled charging load and a cluster model of load formed by the accumulation method are constructed. The simulation results show that the model accurately reflects the spatial and temporal distribution of the charging load of electric vehicles, which lays the foundation for subsequent energy consumption optimization management.

АННОТАЦИЯ

В данной работе предлагается кластерная модель управляемой зарядной нагрузки, основанная на характеристиках поведения пользователей для четырёх типичных сценариев: жилых зон, коммерческих зон, парков и централизованных зарядных станций. Анализируя характеристики сети и характеристики обычной нагрузки в различных сценариях, построена единая модель управляемой зарядной нагрузки, а также кластерная модель нагрузки, сформированная методом накопления. Результаты моделирования показывают, что модель точно отражает пространственное и временное распределение зарядной нагрузки электромобилей, что закладывает основу для последующего управления оптимизацией энергопотребления.

Keywords: electric vehicle, distribution networks, controllable load model, cluster, charging stations, charging scenarios.

Ключевые слова: электромобиль, распределительные сети, модель управляемой нагрузки, кластер, зарядные станции, сценарии зарядки.

INTRODUCTION

Recently, the concept of low-carbon environmental protection and energy conservation has gradually become the mainstream of social development, and the new energy green industry has developed continuously. The electric vehicle (EV) industry will also grow and develop, and in the future, electric vehicles will be used as a new means of transportation to replace traditional fuel vehicles [1].

The share of electric vehicles in the motor vehicle market is also increasing. Large-size electric vehicles have appeared. They have alleviated the situation of fossil energy to a certain extent, but they have also created many problems in the operation of the power grid [2-3]. Since the energy source of electric vehicles is electrical energy, when large-size electric vehicles connected to the power grid are randomly charged, the quality of electricity and the reliability of the distribution network are reduced, which may lead to overload of the power line [4-8]. Therefore, it is necessary to analyze and study the properties of the new load of electric vehicles, and formulate a management strategy. The EV scheduling problem mainly focuses on providing an optimal charging and discharging pattern for EVs in a given time period.

RESEARCH METHODS AND MODELS

To conduct a simulation analysis and modeling of the controlled cluster load, several preparatory stages were carried out. These stages included the characterization of electric vehicle charging scenarios (definition of basic scenarios, description of electrical networks, analysis of typical loads), as well as a mathematical description of the controlled charging load. A more detailed description of the steps is provided below.

1 Characterization of electric vehicle charging scenarios

In this paper, four basic scenarios of EV charging are used for modeling: residential area, commercial area, park and centralized charging station.

The residential area generally refers to a residential area in a city. The load in the charging scene is mainly concentrated in the afternoon and the evening, and is mainly generated by the charging of private cars after work. In most cases, charging is slow.

A commercial district generally refers to a concentrated retail area within a city. The load generated by such a charging area is mainly concentrated in the daytime. Most of the charging activity is based on fast charging.

The park refers to a special environment built by public institutions or enterprises to achieve the purpose of industrial development. The normal charging time is during daytime working hours, the charging activity and power consumption are generally higher in the morning than in the afternoon, and the charging method is slow charging.

Centralized charging stations are specialized charging stations, including large-scale charging stations for buses, large taxis parks, logistics vehicles. The charging behavior is characterized by temporal aggregation, a large part is concentrated at night, and a small part is at noon. Slow charging is mainly used at night, and fast at day.

A detailed discussion of the charging scenarios and the charging types (fast/slow) provides a clear basis for research and data collection on cluster charging, and provides data for analyzing the characteristics of the charging load in different scenarios.

1.1 Grid characteristics of distribution networks in different scenarios

Charging stations are connected to the power grid via a distribution transformer. Depending on the characteristics of the different scenarios and the location of the charging stations, the power levels of the distribution transformers can be vary depending on the charging scenarios. In residential area used one or more public distribution transformers; in business district used multiple public distribution transformers; in park used multiple public distribution transformers or a special distribution one; in centralized charging station used a special distribution transformer.

It can be seen that different charging scenarios have different distribution transformer networks, and different types of distribution transformers also reflect the rules of the actual charging load level. Therefore, the charging load when using dedicated transformers is generally concentrated at night and during the day, with a large difference between the peak and the trough, which demonstrates the phenomenon of night peak and daytime trough. While using public transformers, the charging stations are relatively dispersed, the charging load exists throughout the day, and the overall difference in load amplitude is small.

1.2 Regular Load Characterization for Different Scenarios

Since centralized charging stations are usually fed by separate transformers and the load connected to their distribution transformers is almost entirely charging load and the share of total load is relatively small, only total loads in residential areas, commercial areas and industrial parks were analyzed.

(1) Analysis of residential load characteristics

Residential loads mainly consist of household appliances, including lighting, washing machines, refrigerators, televisions, and other common appliances, as well as some high-energy-consuming household appliances such as air conditioners, electric heaters, electric cookers, and electric water heaters. Lighting power consumption by household appliances varies considerably during the day, but the time difference between loads is small and the simultaneity ratio is relatively high. Figure 1 shows a typical curve based on summer load data in several residential areas.

Figure 1. Typical residential load characteristic curve

(2) Analysis of commercial load characteristics

Commercial loads mainly fall on large commercial buildings such as department stores, entertainment centers, supermarkets, catering establishments and office buildings. The characteristics of commercial and office buildings are as follows: the load of commercial and office buildings shows a pronounced time and season dependence. Commercial loads have a pronounced single-peak and single-trough characteristics, and the difference between the peaks and troughs on the curve is very large. The load curve in Figure 2 is based on load data from several commercial areas. reflecting typical characteristics of a typical commercial load.

Figure 2. Typical commercial load characteristic curve

(3) Analysis of park load characteristics

Park load is the largest industry in terms of electricity consumption, including the following industries: textile, paper, mining, alkali, light industry, petrochemical industry, cement, construction, foundry, coal, machine building, etc. There are many kinds of load industries in the park, and these industries have different energy consumption characteristics, and the energy consumption characteristics of each industry are very different. In general, the park load value is large and relatively stable. The load curve of Figure 3 is fitted by the load data of multiple park areas.

Figure 3. Typical park load characteristic curve

2 Multi-scenario controllable charging load modeling

2.1 Characterization of Electric Vehicle User Behavior

Considering that the driving habits of car owners will not change significantly due to the change in vehicle type, it can be concluded that the driving rules for electric vehicles will be similar to those of traditional cars. Next driving rules were adopted.

(1) Electric bus driving rules

The first bus departs at 5:30 am, and the last one at 11:00 pm. Almost all vehicles are involved in the movement, and the interval between departures is 10 minutes.

(2) Electric taxi driving rules

Taxi drivers are divided into two working classes: a large class and a small class. When the driver is on a large shift, the 24-hour shift is reversed, and the rest is only from 2:00 am to 5:00 pm and from 11:30 am to 2:30 pm during the lunch break. When the driver is on a short shift, the 12-hour shift is used once, and the rest time is relatively short. Rest periods: from 14:00 to 16:00 and from 11:30 to 14:00. At other times, all taxis operate as usual.

(3) Driving Rules for Private Electric Vehicles

Private car owners generally use cars to go to work, return from work, and for leisure. According to the study of the law of electric vehicle behavior, the peak time for the owner to leave work is from 7:30 to 8:30, and the peak time for arriving at work is from 8:00 to 9:00. The peak time for the owner to leave work is from 17:00 to 18:30, and the peak time for staying at home is from 17:30 to 19:00.

2.2 Controlled charging load modeling

In this study, the single model accumulation method is used to obtain a cluster controllable charging load model. Connecting an electric vehicle to the distribution network creates an initial charging load, that is, the initial charging load is the charging load generated by the power grid without regulating the consumption of the electric vehicle itself, and different users have different desires to control the degree of control.

If the starting charging time of the individual electric vehicle is t₀, the original charging time is T₀, the maximum acceptable charging time is T₁, and the original charging power is P₀. The original charging load of individual electric vehicles is calculated as follows:

                                                                  (1)

In the formula: t is in the range of [t₀，t₀+T₀], and P(t) denotes the charging load at time t during this period. The maximum acceptable charging time T₁ of the formula and the controlled preference R_p in the subjective characteristic behavior of the charging user can be calculated by the following formula:

                                                        (2)

In the formula: R_p is controlled preference;  is the controlled duration adjustment coefficient, , this article takes 1.5.

Based on the single charging load model, the charging load of the cluster electric vehicle in the large scene can be further calculated. The original load of the cluster is calculated as follows:

                            (3)

In the formula:  is the original charging load of the electric vehicle cluster at the time of scene t; , ,  and  are the number of buses, taxis, official cars and private cars in the controllable state of charging at time t, respectively; , ,  and  are the original charging power of single bus, taxi, official car and private car at time t, respectively.

The controllable charging load of the electric vehicle cluster is calculated as follows:

                                         (4)

In the formula:  is the controllable charging load of the electric vehicle cluster at the time t; , ,  and  are the controllable charging power of single bus, taxi, official car and private car at time t, respectively.

3 Simulation analysis

Based on the cluster load model, by setting various user characteristic parameters, the subjective behavior of the corresponding users was generated by random sampling, and then the original and controlled charging loads were determined. The simulation parameters are shown below.

Table 1.

The number setting of electric vehicles in different scenarios

Table 2.

Charging user feature parameter settings in different scenarios

According to the daily analysis of users' charging behavior, the following constraint settings are made for the sampling:

① Bus: if it is at night, it is slow charging, and if it is the day, it is fast charging.

② Taxi: if it is at night, it is slow charging; if it is during the day, it is fast charging.

③ Official vehicles: all adopt slow charging.

④ Private cars: in residential areas - slow charging, in business areas – fast.

⑤ All kinds of electric vehicles use constant power charging.

RESULTS AND DISCUSSION

By means of simulation, in accordance with the above scenario settings, the following controlled charging load graphs were obtained.

(a) Residential district

(b) Commercial district

(c) Industrial Park

(d) Centralized charging station

Figure 4. The controllable charging load characteristics of cluster in different scenarios

It can be seen from the above figure that the proposed controlled charging load model can obtain the characteristics of the controlled charging load for one day under the current conditions, taking into account the user's charging needs and combining the distribution of the number of different types of EVs in different scenarios.

In all 4 scenarios considered, the advantage of controlled charging over uncontrolled charging, in the form of a reduction in the amount of charging load, is obvious. In addition, the amplitude of the charging load between peaks and troughs has also noticeably decreased. As a result, minimizing load fluctuations and a smoother charging load curve increase the reliability of the electrical network and increase its efficiency.

CONCLUSION

In this paper, firstly, the characteristics of the networks in residential areas, commercial areas, industrial parks and centralized charging stations were analyzed, and the characteristics of the normal load in the first three scenarios with obvious characteristics of the normal load were analyzed. Then, the interactive operation mode, equipment and control methods of electric vehicles and distribution networks were explained in sequence. Finally, a multi-scenario cluster controllable charging load model was established based on the subjective behavior model based on deep learning, and its practical applicability was verified through simulation examples.

References:

1. Tian, J.; Wang, P.; Zhu, D. Overview of Chinese new energy vehicle industry and policy development // Green Energy and Resources. – 2024, 2(2): 100075.
2. Sun, L.; Teh, J.; Liu, W.; et al. Impact of electric vehicles on power system reliability and related improvements: A review // Electric Power Systems Research. – 2025, 247: 111838.
3. Ramkumar, G.; Kannan, S.; Mohanavel, V.; et al. The Future of Green Mobility: A Review Exploring Renewable Energy Systems Integration in Electric Vehicles // Results in Engineering. – 2025: 105647.
4. Ibrahim, R. A.; Gaber, I. M. Electric vehicles: From charging infrastructure to impacts on utility grid // Power, Energy and Electrical Engineering. IOS Press. – 2024: 326-337.
5. Luan, X.; Guo, Y. Research on EV charging scheduling strategy based on multi-objective optimization // 2024 IEEE 2nd International Conference on Power Science and Technology (ICPST). IEEE. – 2024: 2000-2004.
6. Bhattacharjee, T.; Rajeshwari, M.; Kumar, K. J. Electric Vehicle Charging and its Effects on the Power Distribution Network–A Review // Power Research-A Journal of CPRI. – 2024: 89-98.
7. Wang, S.; Luo, Y.; Yu P, et al. Integrated Coordinated Control of Source–Grid–Load–Storage in Active Distribution Network with Electric Vehicle Integration // Processes. – 2025, 13(5): 1285.
8. Arslanoğlu, İ.; Büyük, M. Electric Vehicle Impacts on Electric Power System: A Literature Survey // 2025 7th International Congress on Human-Computer Interaction, Optimization and Robotic Applications (ICHORA). IEEE. – 2025: 1-7.