A Multi-Scheme Comparison Framework for Ultra-Fast Charging Stations with Active Load Management and Energy Storage Under Grid Capacity Constraints

Abstract

Grid capacity constraints present a prominent challenge in the construction of ultra-fast charging (UFC) stations. Active load management (ALM) and battery energy storage systems (BESSs) are currently two primary countermeasures to address this issue. ALM allows UFC stations to install larger-capacity transformers by utilizing valley capacity margins to meet the peak charging demand during grid valley periods, while BESSs rely more on energy storage batteries to solve the gap between the transformer capacity and charging demand This paper proposes a four-quadrant classification method and defines four types of schemes for UFC stations to address grid capacity constraints: (1) ALM with a minimal BESS (ALM-Smin), (2) ALM with a maximal BESS (ALM-Smax), (3) passive load management (PLM) with a minimal BESS (PLM-Smin), and (4) PLM with a maximal BESS (PLM-Smax). A generalized comparison framework is established as follows: First, daily charging load profiles are simulated based on preset vehicle demand and predefined charger specifications. Next, transformer capacity, BESS capacity, and daily operational profiles are calculated for each scheme. Finally, a comprehensive economic evaluation is performed using the levelized cost of electricity (LCOE) and internal rate of return (IRR). A case study of a typical public UFC station in Tianjin, China, validates the effectiveness of the proposed schemes and comparison framework. A sensitivity analysis explored how grid interconnection costs and BESS costs influence decision boundaries between schemes. The study concludes by highlighting its contributions, limitations, and future research directions.

1. Introduction

With the rapid growth of the electric vehicle market and increasing user demands for charging convenience and energy replenishment efficiency, significant progress has been made in recent years in high-power charging, particularly in the application of ultra-fast charging technology. The construction demand for ultra-fast charging stations is expected to grow rapidly [1]. However, ultra-fast charging stations also place higher requirements on the grid connection capacity. How to solve the grid capacity constraints faced by ultra-fast charging stations has become a key challenge for the widespread deployment of ultra-fast charging stations [2,3].

Ultra-fast charging stations currently adopt two approaches to address capacity constraints. The first involves implementing active load management (ALM) mechanisms or technical solutions. Leveraging the off-peak charging characteristics of ultra-fast charging stations, this approach allows stations to install higher-capacity transformers based on the spare capacity of the connected distribution grid to meet the peak charging demand that usually occurs during grid valley periods. The power output of the transformer is dynamically limited according to the available capacity of the connected distribution network at different times of day, which, on the one hand, improves the service capacity of the station and at the same time helps to improve the utilization efficiency of the distribution network [4,5,6,7,8,9,10,11,12,13,14,15,16]. The second approach involves deploying battery energy storage systems (BESSs) to boost the station’s power supply capacity. Currently, this method is more commonly used under passive load management (PLM). When the transformer’s power supply capacity is insufficient, the BESS enhances the charging power output of ultra-fast charging stations [17,18,19]. Additionally, in regions with significant peak–valley electricity price differentials, some stations are experimenting with larger-capacity battery storage to reduce the overall power supply costs [20,21].

In addressing grid capacity constraints for ultra-fast charging stations using the ALM approach, Weisbach et al. [6] proposed a DC charging station with ALM that was developed to react to fluctuating charge current acceptance to avoid underutilization. Ramaschi et al. [7] pointed out that using ALM schemes is effective at improving the service capacity of charging stations, which can shorten the charging time of users, increase the satisfaction rate of electric vehicle users, and improve the profitability of charging point operators. In [8], the communication protocol standard is presented as a communication protocol standard that provides an important support for supporting the application of ALM and the development of communication control schemes between grid operators and charging stations. Bayram et al. [22] increased the capacity limit in the low valley hours through ALM, and at the same time, with its ALM strategy, it achieves a smooth, flexible, and efficient connection between distributed power sources and the grid, thus optimizing the operational performance of the grid and reducing the impact of distributed power access on the grid. Nayak et al. [23] designed a dynamic load management system that uses load management strategies to dynamically adjust battery power of electric vehicles (EVs); effectively coordinating renewable energy sources, storage systems, the grid, and charging stations; and responding quickly when new peaks in the load profile occur.

In addressing grid connection capacity constraints for charging stations using BESSs, Zhang et al. [24] examined the impact of distribution network constraints on the charging station transformer capacity and proposed the use of energy storage technologies to reduce the transformer capacity requirements and enhance the load management capabilities. He et al. [25] and Han et al. [26] further develop a capacity optimization method based on PV–storage hybrid systems, utilizing energy storage for peak shaving and valley filling to mitigate grid impacts. Additionally, Zhang et al. [27] account for seasonal fluctuations in the EV charging demand and battery lifespan degradation, optimizing the energy storage configuration for charging stations. Gui et al. [28] and Li et al. [29] propose power demand suppression methods based on charging optimization and energy storage allocation, which demonstrate the effectiveness of energy storage systems in reducing load fluctuations in the power grid.

In the context of the economic comparison of different construction schemes for charging stations, the Levelized Cost of Energy (LCOE) and the Internal Rate of Return (IRR) are widely used two indicators and methods. LCOE allocates the project’s investment, operation, and maintenance costs over its entire lifecycle to each unit of electricity, thereby objectively quantifying the cost competitiveness of the project. This makes it particularly suitable for evaluating the economic viability of multiple alternatives [30,31,32,33]. Lanz et al. [32] systematically categorized charging equipment, collected extensive market data on device costs, and employed LCOE to model charging costs across 30 European countries and 13 different private passenger charging options. Additionally, Deng et al. [33] analyzed the economic impacts generated by photovoltaic (PV) and energy storage (ES) equipment constructed in DC fast-charging stations separately for PVs and ES using LCOE. IRR, calculated as the discount rate that yields a net present value (NPV) of zero, reflects the investment return of a charging station under given electricity prices and load scenarios. It allows for a direct comparison between the IRR and a company’s benchmark return rate, thereby providing clear support for investment decisions on individual projects [34,35,36,37]. Ghosh et al. [35] used IRR to evaluate the feasibility of retrofitting existing gas stations with fast-charging infrastructure. However, IRR is not suitable for comparing multiple alternatives with significant differences in investment scale. Sun et al. [36,37] primarily relied on IRR and NPV-based return analyses to assess the economic viability of PV–storage charging stations and battery operation models. While these studies aim to maximize investment returns, they often overlook the cost per unit of electricity, making it challenging to reflect long-term operational cost competitiveness. Therefore, when comparing multiple alternatives, it is essential to leverage the strengths of both LCOE and IRR. LCOE should be prioritized for cross-scheme cost competitiveness comparisons, while IRR can then be used to evaluate the investment returns of economically superior schemes.

Overall, significant research has been conducted on addressing charging station capacity constraints through ALM and BESSs. Numerous achievements have been made in developing power allocation strategies for charging stations under capacity limitations and optimization algorithms for energy storage, demonstrating effectiveness in improving charging station economics and enhancing charging service capabilities. However, few studies have explored the combined schemes of ALM and BESS solutions. To address the grid connection capacity constraints faced by ultra-fast charging stations, designing feasible, combined solutions of ALM and BESSs and establishing a multi-scheme evaluation framework are necessary. This will support charging infrastructure investors in developing specific project construction plans, while also providing a decision-making basis for government departments to formulate relevant support and regulatory policies.

Therefore, a four-quadrant classification of scheme combinations for ultra-fast charging stations to solve grid capacity constraints is proposed, defining four types of schemes based on ALM and the BESS: ALM with a maximum storage configuration (ALM-Smax), ALM with a minimum storage configuration (ALM-Smin), PLM with a maximum storage configuration (PLM-Smax), and PLM with a minimum storage configuration (PLM-Smin). A generalized comparison framework is further proposed for the four feasible schemes, in which the transformer capacity, BESS capacity, and typical operating profiles of the four schemes under grid capacity constraints are calculated through a typical daily charging load demand simulation, and a comprehensive economic comparison is carried out based on the two indicators of LCOE and IRR.

The remainder of this paper is organized as follows: Section 2 introduces the ALM-BESS four-quadrant classification framework and generalized comparison framework for solving grid capacity constraints at ultra-fast charging stations. Section 3 introduces detailed methods for the configurations of transformer capacity and BESS capacity and the comprehensive economic comparison method. Section 4 presents the results of a case study of a typical ultra-fast charging station in Tianjin, China. Section 5 gives the main conclusions and recommendations for further research.

2. ALM-BESS Four-Quadrant Classification Framework and Generalized Comparison Framework

2.1. ALM-BESS Four-Quadrant Classification Framework with Definitions of the Four Schemes

When facing the grid-connected capacity constraints, ultra-fast charging stations can form different schemes from two dimensions: ALM and BESS. In this paper, a four-quadrant classification framework of ALM-BESS is proposed, as shown in Figure 1. According to this classification framework, four types of schemes can be defined, which are PLM with a minimum BESS (PLM-Smin), PLM with a maximum BESS (PLM-Smax), ALM with a minimum BESS (ALM-Smin), and ALM with a maximum BESS (ALM-Smax).

2.1.1. PLM-Smin Scheme

The UFC station is connected to the grid in a passive load management mode and the grid operator does not apply capacity management or communication control measures to the charging station loads; so, the main way to ensure the safety of grid operation is through the capacity constraints of the transformer accessed by the station, which is not allowed to exceed the available capacity of the accessed distribution network during peak hours. Therefore, when facing grid capacity constraints, introducing a BESS in PLM mode and reasonably configuring its energy and power capacities are necessary to ensure that the maximum charging power demand of the vehicles in the stations can be met under the limited transformer capacity. Since the peak charging loads of vehicles at fast charging stations usually occur in the low valley period of the grid under the guidance of peak and valley electricity price, it is easy to cause the BESS to be discharged in the low valley period of the grid (corresponding to the low valley electricity price), and the economic benefit of the BESS discharging is depressed.

2.1.2. PLM-Smax Scheme

On the basis of PLM-Smin, this scheme further aims to maximize the value of the BESS peak–valley arbitrage by further increasing the scale of the BESS, and realizing the peak–valley arbitrage by discharging during the peak hours of the grid load (corresponding to the peak electricity price) and recharging during the valley hours of the grid load (corresponding to the valley electricity price). In this case, the scale of the BESS configuration will increase significantly compared to PLM-Smin, and in the actual implementation of the project, taking into account the land requirement of the energy storage facilities and the higher requirements for fire risk prevention after the increase in the scale of the BESS are necessary.

2.1.3. ALM-Smin Scheme

The ALM scheme achieves dynamic management of the capacity of the site by introducing active load management strategies, including limiting the output of the transformer capacity at peak and flat periods by the users themselves according to the grid connection agreement, or establishing a direct communication control path between the grid operator and the station. So, the station can be allowed to connect to a transformer with a larger capacity that exceeds the available capacity of the distribution network during peak hours. Under the ALM scheme, considering that under the guidance of peak and valley electricity prices, the vehicle charging load peak of fast charging stations usually occurs in the valley hours of the power grid load, at which time the vehicle charging demand can be met more directly through the transformer capacity, but due to the fact that there may still be a gap in the dynamically available capacity of the transformer in different time periods, configuring a certain amount of capacity of the BESS is necessary to ensure that vehicle charging demand in various time periods. But, the capacity demand of the BESS will be reduced significantly compared to the PLM-Smin scheme.

2.1.4. ALM-Smax Scheme

This scheme is based on ALM-Smin and further aims to maximize the value of the BESS peak–valley arbitrage by further increasing the scale of the BESS. Generally speaking, since the capacity of the transformer has been significantly increased compared to PLM mode, the amount of electricity and power space that can be engaged in grid peak–valley arbitrage in the ALM-Smax scheme is also significantly larger than that in the PLM-Smax scheme, which means that the scale of the BESS configured in the scheme also significantly exceeds that in the PLM-Smax scheme. Therefore, in the actual implementation of the project, ALM-Smax needs to consider the space requirements and fire risk management measures for energy storage facilities.

2.2. Generalized Comparison Framework for the Four Schemes

Based on the four-quadrant classification framework proposed above, when the grid operator provides an ALM interconnection solution, the investor of the ultra-fast charging station construction needs to compare the four types of schemes when solving the problem of grid capacity constraints in order to determine the solution that is most suitable for its station construction needs and has the best cost competitiveness and return on investment. Therefore, for the ALM-BESS four-quadrant classification framework, a generalized comparison framework for the four schemes is further proposed, as shown in Figure 2.

The generalized comparison framework for the four types of schemes consists of three main steps. In the first step, based on the pre-set vehicle service scale and charging pile configuration parameters, the Monte Carlo method is used to simulate the vehicle charging process at the station to form a typical daily charging load curve. In the second step, the capacity configurations of the transformer and the BESS under the four schemes are calculated, as well as the simulated transformer and the BESS under the typical daily operating conditions. In the third step, the initial investment cost, annual operation and maintenance (O&M) cost, and charging revenue are calculated, and the LCOE and IRR of each scheme are derived. The LCOE is first used to select alternatives among the four schemes, and then the IRR is used to evaluate whether the level of return on the investment of the alternatives meets the expectations in order to make the final scheme decision.

3. Configurations of the Transformer and BESS and Comprehensive Economic Evaluation

3.1. Typical Daily Charging Load Simulation of UFC Stations

Concerning the total amount of vehicles served in the station on average per day and the given charger configuration scheme, the hourly user arrival probability distribution curve is first constructed, and then the Monte Carlo method is used to synthesize the typical daily load curve of the station through multiple rounds of stochastic simulation. In each round of the Monte Carlo simulation, the charging process of vehicles at the station is iteratively simulated with a time step of minutes: at any moment t, judge if the vehicles are fully charged at the moment t − 1, and then obtain the connection status of each vehicle with the chargers and the corresponding charging power. According to the state at t − 1, determine whether there is a new arrival at the moment t. And, update the occupancy of the chargers and the charging power, and calculate the charging power of the charging station. The total pow of station $P_t$ is the sum of the power of all the chargers $P_{i,t}$ The formula is shown in Equation (1) and the flow figure is shown in Figure 3.

$$P_t={\displaystyle \sum _{i=1}^n}{P}_{i,t}$$

(1)

where $P_t$ is the total charging power of the station at t, n is the quantity of the chargers of the station, $P_{i,t}$ is the power of i-th charger at the moment t.

3.2. Transformer and BESS Capacity Configurations and Typical Daily Operating Profile

This section introduces how to configure the transformer and BESS capacities and obtain the typical daily operating profile of the transformer and BESS.

3.2.1. Transformer Capacity Configuration and Typical Daily Operating Profile

• PLM-Smin scheme and PLM-Smax scheme

For the PLM scheme, in the case where the grid capacity is limited to meet the maximum charging demand, the transformer capacity is configured by the upper limit of the available capacity of the grid during peak hours, as shown in Equation (2).

$$C=Q_p$$

(2)

where C is the configuration capacity of the transformer and $Q_p$ is the the available capacity of the grid during peak hours.

2.

ALM-Smin scheme and ALM-Smax scheme

Firstly, the maximum charging load of the station in each time period is calculated as the maximum peak time load P_p,max, the maximum flat time load P_f,max, and the maximum valley time load P_v,max. Then, the load peaks of each period are compared with the transformer’s available capacity limits Q_p, Q_f, and Q_v, taking the smaller value as the permitted transformer output capacity for each period. The maximum value is selected among the permitted output capacities of the three periods as the transformer configuration capacity. To enhance system adaptability and operational safety, a reasonable margin coefficient can be introduced to appropriately increase the transformer configuration capacity. The calculation formulas for the transformer configuration capacity under ALM mode are shown in Equations (3)–(6), and the calculation process is illustrated in Figure 4.

$$\left\{\begin{array}{c}{C}_v=min(P_{v,max},Q_v)\\ C_f=min(P_{f,max},Q_f)\\ C_p=min(P_{p,max},Q_p)\end{array}\right.$$

(3)

$$C_x=max(C_v,C_f{,C}_p)$$

(4)

$$C_y=max(Q_v,Q_f{,Q}_p)$$

(5)

$$C=min\left(\right(1+r_y)C_x,C_y)$$

(6)

where $r_y$ is the margin coefficient, $C_x$ is the maximum value of the permitted output capacity of the transformer for different time periods, and $C_y$ is the maximum value of the upper limit of available capacity for different time periods.

3.2.2. BESS Capacity Configuration and Typical Daily Operating Profile

• PLM-Smin scheme and PLM-Smax scheme

First, compare the minute-level load curve L_t with the transformer capacity upper limit C to obtain the minute-level capacity power deficit e_t. Cumulatively sum these deficits to derive the total deficit energy E, which serves as the energy capacity value of PLM-Smin. Meanwhile, the maximum power deficit e_t observed during the calculation is designated as the power capacity value of PLM-Smin. To achieve maximal peak–valley arbitrage, calculate the load reduction value sh_t for each minute of the peak period based on maximizing peak load reduction. Cumulatively sum these reductions to obtain the total peak-shaving energy Sh, which is then combined with E to form the energy capacity value of PLM-Smax. Finally, take the maximum value between the sum of reduction values sh_t and limit violation value e_t as the power capacity value of PLM-Smax.

The calculation formulas are shown in Equations (7)–(10), and the calculation process is illustrated in Figure 5. In the calculation process, the peak-shaving ratio r is used to switch between PLM-Smin and PLM-Smax computations:

• In the PLM-Smin scheme, r = 0;
• In the PLM-Smax scheme, the peak-shaving ratio r is determined by back-calculating the reducible peak load proportion based on the maximum available energy storage charging capacity during valley periods, which defines the corresponding peak-shaving ratio.

$$Q_1=E/60\eta$$

(7)

$$P_1=e_{t,max}$$

(8)

$$Q_2=(E+Sh)/60\eta$$

(9)

$$P_2={max(e}_t+{sh}_t)$$

(10)

where $Q_1$ is the BESS energy capacity of the PLM-Smin scheme, $\eta$ is the discharge efficiency, $P_1$ is the BESS power capacity of the PLM-Smin scheme, $Q_2$ is the BESS energy capacity of the PLM-Smax scheme, and $P_2$ is the BESS power capacity of the PLM-Smax scheme.

2.

ALM-Smin scheme and ALM-Smax scheme

For the ALM-Smin scheme, the following implementation steps should be strictly followed: First, based on the predicted load curve shown in Figure 3 and grid connection capacity limits during different time periods, the transformer’s available capacity limits must be established for specific intervals—designated as $C_p$, $C_f$, and $C_v$ for peak, flat, and valley periods, respectively. A point-by-point comparison between minute-level predicted load L_t and the corresponding period capacity limit generates the exceedance power value $e_t$ for each minute. The total exceedance value E equals the sum of all these exceedance values $e_t$. Concurrently, the maximum power gap e_t identified during computation serves as the power capacity value for PLM-Smin. To achieve maximal peak–valley arbitrage, peak load reduction values sh_t for each minute are calculated based on the maximum peak load reduction requirements, with the accumulated total peak shaving energy Sh being superimposed with E to form the energy capacity value for PLM-Smax. The maximum value between the sum of reduction values sh_t and exceedance values e_t determines the power capacity value for PLM-Smax.

The calculation formulas are shown in Equations (7)–(10), and the calculation process is illustrated in Figure 6. In the calculation process, the peak-shaving ratio r is used to switch between ALM-Smin and ALM-Smax computations:

• In the ALM-Smin scheme, r = 0;
• In the ALM-Smax scheme, r = 1.

3.3. Comprehensive Economic Selection Methodology

The section compares the construction costs and O&M costs between the four schemes.

3.3.1. Cost–Benefit Components of the Four Schemes

The cost structure of ultra-fast charging stations primarily consists of initial investment costs, O&M costs, electricity costs, rental expenses, and other associated expenditures. Revenue streams are derived from user payments, including electricity fee revenue and service fee revenue. Detailed breakdowns of these cost and income components are illustrated in Figure 7. Under different operational modes, the compositions and proportions of these cost items exhibit significant variations, as detailed below.

• Initial Investment Cost

During the initial phase, station investors must allocate fixed asset investments related to station construction, including grid-connection facilities such as transformers, chargers, energy storage systems, and associated construction and equipment costs. The corresponding computational models are detailed in Equation (11).

$$C_0=C_{tran}+C_{stack}+C_{cg}+C_{sto}$$

(11)

where $C_{tran}$ is the construction and equipment costs for transformer, $C_{stack}$ is the investment in charging stack infrastructure and associated equipment, $C_{cg}$ is charger installation and hardware expenditures, and $C_{sto}$ is BESS deployment and component costs, which are explained in Appendix A.

In terms of the transformer investment, the capacity configuration of the PLM scheme is determined by the available capacity during peak periods of the connected distribution network, whereas the ALM mode optimizes its configuration through a comprehensive consideration of capacity margin during off-peak periods, user charging demand, and peak-shaving requirements, resulting in relatively higher transformer investment costs under the ALM mode. Regarding the energy storage system configuration, both the PLM-Smin and ALM-Smin schemes primarily aim to address limit violation issues, thus requiring relatively smaller energy storage capacities. However, due to the ALM mode’s larger transformer capacity compared to the PLM-Smin scheme, its demand for energy storage is further reduced, achieving the minimal required energy storage configuration. In contrast, the PLM-Smax and ALM-Smax schemes not only need to satisfy limit constraints but also maximize peak–valley arbitrage. Among these, the ALM-Smax scheme demonstrates greater potential for peak–valley arbitrage owing to its superior transformer capacity, consequently demanding the largest energy storage system and incurring the highest corresponding investment costs. As the construction scale of charging piles and charging stacks is determined based on projected planning, the investment costs for these components remain consistent across all four modes.

2.

Annual O&M costs and rental expenses

The O&M costs of ultra-fast charging stations primarily refer to expenses incurred during annual operation and maintenance processes. The total O&M costs of ultra-fast charging stations mainly consist of the transformer, charging pile, power stack, and energy storage system maintenance costs. The calculation methodology for these O&M costs is detailed in Equations (12) and (13).

$$C_{OM}=C_{OM,tran}+C_{OM,cg}+C_{OM,stack}+C_{OM,sto}$$

(12)

$$C_{OM,sto}=S_{ic}·C_{sto}$$

(13)

where $C_{OM}$ is the station O&M cost, $C_{OM,tran}$ is the transformer O&M cost, $C_{OM,cg}$ is the charger O&M cost, $C_{OM,stack}$ is the charging stack maintenance cost, $C_{OM,sto}$ is the BESS O&M cost, and $S_{ic}$ the annual O&M rate for BESS, which are explained in Appendix A.

Rental expenses are primarily a cost incurred by the site investor when leasing space, as shown in Equation (14).

$$C_{rent}={RT}_s·N$$

(14)

where ${RT}_s$ is the annual rental price of parking spaces and N is the quantity of parking spaces.

The O&M costs of transformers are positively correlated with their capacity—larger capacities result in higher O&M expenses. For energy storage systems, the O&M costs depend on both the storage investment and the O&M rate, with higher values for either factor leading to increased costs. Since the quantities of charging piles, charging stacks, and parking spaces remain identical across the four operational modes, their O&M costs and rental fees remain consistent across different modes.

3.

Electricity costs

The electricity cost of the station comprises the energy charge and capacity charge. The energy charge primarily stems from the station’s electricity procurement expenditures during daily operations, including power purchases for user charging and energy storage charging, as shown in Equation (15).

$$P_{co}=p_{e,v}·E_v+p_{e,f}·E_f+p_{e,p}·E_p$$

(15)

where $p_{e,v}$, $p_{e,f}$, and $p_{e,p}$ are the electricity price during peak, flat, and valley periods, respectively, and the unit is RMB. $E_v$, $E_f$, and $E_p$ are the electricity consumption of the station during peak, flat, and valley periods, respectively, and the unit is kWh.

Peak–valley arbitrage through a BESS can effectively reduce the station’s electricity costs. Under identical user demands across the four operational modes, the ALM-Smax mode demonstrates the lowest station electricity costs due to its maximum peak–valley arbitrage capacity, while the PLM-Smin mode results in the highest electricity expenses owing to its minimal arbitrage scale.

The capacity charge is primarily calculated based on the transformer capacity, as shown in Equation (16).

$$P_{ca}=p_{ca}·C_m$$

(16)

where $P_{ca}$ is the capacity charge, which means the part of electricity price determined by the transformer capacity, and the unit is RMB. $p_{ca}$ is the unit cost of the capacity charge, and the unit is kVA/RMB. $C_m$ is the transformer capacity.

The capacity charge depends on the method used to determine the transformer capacity. If the determination is based on the configured capacity C, then the ALM’s capacity charge cost will be higher than that of PLM. However, if the determination is based on the dynamic available capacity Q_p during peak hours, the capacity charge cost of ALM can be on par with that of PLM.

• Station revenue

The annual revenues of the UFC station include service charges and electricity paid by users, as shown in Equations (17) and (18).

$$Y_{year}=Y_s+Y_e$$

(17)

$$Y_s=p_s·E_t$$

(18)

where $Y_{year}$ is the annual revenue of the UFC station, and the unit is RMB. $Y_s$ is the service fee paid by users based on the vehicle charging capacity, and the unit is RMB. $Y_e$ is the electricity charge collected by the station based on the vehicle charging capacity, and the unit is RMB. Given that the vehicle charging capacity remain consistent across the four schemes, their corresponding revenues do not change.

In summary, the costs and revenues of the four scenarios are characterized in

Table 1. Distribution characteristics of costs and revenues under the four schemes.

Category		PLM-Smin	PLM-Smax	ALM-Smin	ALM-Smax
Initial Investment	Grid connection	Low		High
	BESS	Lower	Higher	Lowest	Highest
	chargers	Same
	Stacks	Same
Annual O&M Costs	Transformer	Low		High
	BESS	Lower	Higher	Lowest	Highest
	Power stack and chargers	Same
Rent	Parking Space Rent	Same
Electricity Costs	Electricity Cost	Highest	Lower	Higher	Lowest
	Capacity Cost	Lower		Higher/Equal
Revenue	Electricity Fee	Same
	Service Fee	Same

.

3.3.2. Economic Evaluation Process Based on LCOE and IRR

LCOE Calculation Model

The Levelized Cost of Energy (LCOE), proposed by the National Renewable Energy Laboratory (NREL) in the United States, is a metric that calculates the cost of energy by levelizing the total lifecycle costs and energy generation of a project. It is used to compare the costs of different types of power generation projects on a consistent basis, as shown in Equation (19).

$$C_{LCOE}={\displaystyle \frac{I_0+{\sum }_{t=1}^n\frac{A_t}{{(1+i)}^t}}{{\sum }_{t=1}^n\frac{M_t}{{(1+i)}^t}}}$$

(19)

where $C_{LCOE}$ represents the LCOE; $I_0$ denotes the initial construction investment of the project; $n$ is the project lifespan; $A_t$ refers to the operating costs in year t; $M_t$ indicates the energy generation in year t; and $i$ is the expected rate of return of the project.

The LCOE can be used to calculate the costs of charging stations under different schemes, as shown in Equation (20).

$$C_{LCOE}={\displaystyle \frac{C_0-\frac{V_R}{{(1+r)}^n}+{\sum }_{t=1}^n\frac{C_t}{{(1+r)}^t}}{{\sum }_{t=1}^n\frac{E_t}{{(1+r)}^t}}}$$

(20)

where $C_0$ represents the initial investment of the charging station; $V_R$ denotes the residual value of fixed assets at the end of the project’s operational period; r is the discount rate; $C_t$ refers to the total costs incurred in year t, which is the sum of operation and maintenance (O&M) costs, electricity costs, and rental costs; and $E_t$ represents the electricity sales in year t.

The LCOE is used to compare the economic performance of different schemes on a consistent basis. It reasonably eliminates the differences in initial investment and electricity sales across various charging station schemes. By calculating the ratio of total costs to total energy generation over a specific operational period for each solution, LCOE comprehensively reflects multiple economic factors, such as investment costs, O&M costs, and capital payback periods. This provides a standardized cost metric, enabling decision-makers to effectively compare and evaluate different schemes.

IRR Calculation Model

For the alternative scheme selected based on LCOE in Section LCOE Calculation Model, it is still necessary to evaluate its investment return level through the IRR, which represents the discount rate at which the total present value of capital inflows equals the total present value of capital outflows. The calculation formulas for IRR are shown in Equations (21)–(22).

$$\mathrm{N}\mathrm{P}\mathrm{V}={\sum }_{t=0}^N{\displaystyle \frac{{CF}_t}{{(1+IRR)}^t}}-C_0=0$$

(21)

$${CF}_t=\left\{\begin{array}{c}{C}_0,t=0\\ {RE}_s-{EX}_s,1\le t\le T\end{array}\right.$$

(22)

where NPV represents the net present value; ${CF}_t$ denotes the annual net cash flow of the project at a specific time stage; ${RE}_s$ is the annual revenue of the charging station; and ${EX}_s$ indicates the sum of the annual O&M costs, electricity costs, and rental costs of the charging station.

The IRR helps assess the investment feasibility and capital return of the scheme selected based on the LCOE, thereby providing more comprehensive support for investment decision-making.

Comprehensive Comparison and Selection Process

First, the LCOE and IRR of the four schemes are calculated. Then, the scheme with the lowest LCOE is selected as the optimal alternative for the horizontal comparison, and its IRR is further evaluated against the benchmark investment threshold A. If the IRR meets the requirement, the scheme is identified as the optimal solution and approved for implementation; otherwise, the investment plan is terminated, as shown in Figure 8.

4. Case Study Analysis

4.1. Case Design and Key Parameters

This paper conducts a survey of public UFC stations for electric vehicles in Tianjin, China. In response to the demand for the development of ultra-fast charging, a UFC station with a maximum charging service capacity demand of around 1.2 MW is designed as the case study object. The relevant parameters of the case station are set from four aspects: vehicle charging demand, configuration of chargers at the station, grid connection capacity constraints and peak–valley time-of-use electricity prices, and investment and operation and maintenance (O&M) costs of the UFC station.

4.1.1. Vehicle Charging Load Parameters

Based on the survey results, the designed case charging station is assumed to serve an average of 369 vehicles per day, with the proportion of ultra-fast charging vehicles set at 30% to simulate the impact of their increasing adoption. Considering that the charging process is significantly influenced by the vehicle’s terminal State of Charge (SOC) and that users’ initial SOC levels exhibit considerable uncertainty, this study simplifies the modeling of user charging behavior and power characteristics by assuming a uniform SOC range between 20% and 80%. The battery capacity and charging power rate are determined based on survey data from mainstream battery electric vehicles in Tianjin. The parameter settings for vehicle charging demand in the case study are summarized in

Table 2. Case station vehicle service parameters.

Parameters	Notations	Values
Average Daily Number of Vehicles Served	$N_{ev}$	369
Proportion of Fast-Charging Vehicles (%)	${PR}_{fv}$	70%
Proportion of Ultra-Fast-Charging Vehicles (%)	${PR}_{sv}$	30%
Max Charging Power of Slow-Charging Vehicles (kW)	$P_{sl,ev}$	60
Max Charging Power of Fast-Charging Vehicles (kW)	$P_{f,ev}$	180
Max Charging Power of Ultra-Fast-Charging Vehicles (kW)	$P_{s,ev}$	360
SOC Range (%)	$∆\mathrm{s}\mathrm{o}\mathrm{c}$	$\left[20\mathrm{\%},80\mathrm{\%}\right]$
Battery Capacity (kWh)	$\mathrm{B}\mathrm{C}$	67

.

4.1.2. Chargers’ Configuration Parameters

Based on a field investigation of typical fast-charging stations in Tianjin, the case station is configured with 10 charging piles to accommodate both ultra-fast charging and conventional fast charging demands. The charging piles are allocated across three power levels—120 kW, 200 kW, and 360 kW (ultra-fast charging)—with specific proportions. The overall scale and power configuration of the charging piles at the case station are summarized in

Table 3. Case station configuration parameters.

Parameters	Notations	Values
Number of Chargers	$N$	10
Fast Charger Power (kW)	$P_{oc-1}$	120
Fast Charger Power (kW)	$P_{oc-2}$	200
Ultra-Fast Charger Power (kW)	$P_{sc}$	360
Number of 120 kW Fast Chargers	$N_{oc-1}$	4
Number of 200 kW Fast Chargers	$N_{oc-2}$	3
Number of 360 kW Ultra-Fast Chargers	$N_{sc}$	3

.

4.1.3. Grid-Connection Capacity and Time-of-Use Electricity Price Parameters

Based on the situation of typical “hand-in-hand” distribution lines with high capacity utilization rates in Tianjin, the accessible capacity parameters for the case study station are set. The rated capacity of the 10 kV distribution cable line to be connected to the case station is set at 5 MW, and it is structured as a "hand-in-hand" configuration with another 10 kV line. Both lines need to have the ability to transfer load under N − 1 single-line fault conditions. Therefore, the actual maximum accessible capacity of the 10 kV line proposed for the case ultra-fast charging station is considered as 2.5 MW, with the existing peak load on the line set at 1.87 MW, leaving approximately 630 kW of available capacity. Based on the estimated 40% peak–valley load difference for similar typical distribution lines, the average and off-peak loads for this line are set at 1.6 MW and 1.12 MW, respectively. As a result, different available capacity limits are set for the ultra-fast charging station in peak, off-peak, and valley periods, with the peak capacity not exceeding 630 kW, average period capacity not exceeding 900 kW, and off-peak period capacity not exceeding 1380 kW, as shown in

Table 4. Grid capacity constraints.

Parameters	Notations	Values
Peak Period Capacity Limit (kW)	$Q_p$	630
Flat Period Capacity Limit (kW)	$Q_f$	900
Valley Period Capacity Limit (kW)	$Q_v$	1380

.

Regarding the peak–valley electricity price setting, this study refers to the time-of-use periods and electricity price levels for large industrial electricity users (below 1 kV) in the April 2025 Tianjin electricity price table for agent-purchased electricity [38]. The specific electricity prices and corresponding time periods are shown in

Table 5. Time-of-use electricity tariffs and period division.

Parameters	Notations	Values
Peak period energy price (RMB/kWh)	$p_p$	1.0
Flat period energy price (RMB/kWh)	$p_f$	0.7
Valley period energy price (RMB/kWh)	$p_v$	0.4
Peak period	$T_p$	9:00–12:00, 16:00–21:00
Flat period	$T_f$	7:00–9:00, 12:00–16:00, 21:00–23:00
Valley period	$T_v$	23:00–7:00

. According to China’s current charging infrastructure support policies, capacity charges for charging facilities are not considered at this stage [39].

4.1.4. UFC Station Investment, O&M Costs, and Revenue Parameters

In determining the boundary parameters for the investment and operational costs of the case UFC station, this study investigates the construction and operational expenses of mainstream charging and battery swapping stations currently on the market. Unit cost information for key equipment—including chargers, power stacks, and transformers—was obtained [40]. For the BESS, lithium–iron phosphate batteries were selected due to their high cycle life and safety performance. The economic evaluation takes into account the capacity degradation characteristics, typical cycle life, and relevant performance parameters [41]. Detailed settings of key economic parameters are presented in

Table 6. Investment and operation and maintenance cost parameters of the case station.

Parameters	Notations	Values
120 kW Charger Equipment Cost (10,000 RMB/unit)	$E_{oc-1}$	1.2
200 kW Charger Equipment Cost (10,000 RMB/unit)	$E_{oc-2}$	1.5
360 kW Supercharger Equipment Cost (10,000 RMB/unit)	$E_{sc}$	3
Charger Construction Cost (RMB/unit)	$C_{oc}$	14,000
Supercharger Gun Construction Cost (RMB/unit)	$C_{sc}$	19,000
Standard Gun Maintenance Cost (RMB/Charger/year)	$M_{oc}$	400
Supercharger Gun Maintenance Cost (RMB/Charger/year)	$M_{sc}$	1500
Power Stack Equipment Cost (RMB/kW)	$E_{st}$	400
Power Stack Construction Cost (RMB/unit)	$C_{st}$	40,000
Transformer Construction Cost Coefficient	$T_{ic}$	0.3
Grid Connection Cost (RMB/kVA)	$E_{ic}$	1000
Power Stack and Transformer Maintenance Cost (RMB/unit/year)	$M_{T,st}$	20,000
Parking Space Rent (10,000 RMB/unit/year)	${RT}_s$	1.2
Residual Value Rate of Chargers and Power Equipment	$RE$	0.2
Station Service Life	$SL$	10
BESS Energy Capacity Cost (RMB/kWh)	${\delta }_E$	1500
BESS Power Capacity Cost (RMB/kW)	${\delta }_P$	300
BESS Annual O&M Cost Ratio Coefficient	$S_{ic}$	4%
BESS Cycle Life (cycles)	$L_s$	5000
BESS Average Annual Capacity Degradation Rate	$R_{de}$	2%
BESS Battery Residual Value Rate	${RE}_s$	0.2
Service Fee Unit Price (RMB/kWh)	$p_{se}$	0.3

.

4.2. Configuration and Operation Simulation

4.2.1. Typical Daily Charging Load Demand Modeling

By researching the user arrival distribution patterns at typical stations under the current peak–valley time-of-use electricity pricing policy in Tianjin, a user arrival time distribution curve was constructed, as shown in Appendix B.

Based on the process and formulas outlined in Section 3.1, and using the vehicle and charger configuration parameters set in Section 4.1, a total of 365 Monte Carlo simulation runs were conducted, and the results were averaged to obtain the typical daily charging load demand curve for the case UFC Station, as shown in Figure 9. The maximum load peaks for different time periods are shown in

Table 7. Maximum loads at different periods.

	Peak Period	Peak Period	Valley Period
Maximum Load (kW)	655	714	1162

. The peak charging load for the station occurs during the low electricity price period, with a maximum value of 1162 kW. The maximum loads during the flat and peak periods are 714 kW and 655 kW, respectively, which are significantly lower than the peak value during the valley period. This load distribution indicates a significant peak-shaving characteristic in the charging load.

4.2.2. Configurations of the Four Schemes

Due to the limitations of the grid connection conditions, the capacity margin of the station during peak periods is only 630 kW. If the station is only allowed to install a 630 kVA transformer under the traditional PLM scheme, based on the predicted charging load demand, the peak charging load could reach 1162 kW, resulting in a significant gap in grid capacity. This would lead to users facing waiting times or even abandoning charging services during peak periods, adversely affecting the overall user experience. Based on the charging load demand curve obtained from Section 3.2 and Section 4.2.1, the configuration results for the transformer capacity and BESS capacity of the UFC station in the four schemes are calculated, with the specific configuration parameters detailed in

Table 8. Configurations of the transformer and BESS under the four schemes.

Scheme	Transformer Capacity (kVA)	BESS Energy Capacity (kWh)	BESS Power Capacity (kW)
PLM-Smin	630	1070	600
PLM-Smax	630	1740	600
ALM-Smin	1280	10	50
ALM-Smax	1280	3940	1030

.

As shown, for the two minimal BESS configurations, PLM-Smin and ALM-Smin, where the BESS only supplements the transformer capacity or the available capacity limits and the charging demand gap for each period, the maximum charging demand of the station under peak–valley pricing occurs during the grid valley period. The ALM model allows for the installation of a larger transformer and utilizes the low valley capacity margin of the distribution network. Therefore, the BESS capacity required for the ALM-Smin scheme is significantly lower than that of the PLM-Smin scheme. Conversely, when the BESS is used not only to supplement the power supply capacity gap but also to maximize energy storage configuration to optimize peak–valley arbitrage revenue, the ALM-Smax scheme has a larger transformer capacity, thus providing greater flexibility in regulating peak–valley arbitrage energy. Consequently, the BESS capacity configured for ALM-Smax is significantly larger than that for PLM-Smax.

4.2.3. Simulation of Typical Daily Operating Profiles for Transformers and BESSs

Based on the user load demand forecasting results and the station configuration parameters under the different strategies, and in accordance with the simulation process in Section 3.2, the typical daily load curve in Section 4.2.1, and the station configuration in Section 4.2.2, simulations were carried out for the transformers and BESS under the four schemes over the full period of a typical day. In terms of energy storage system charging strategy planning, grid load characteristics and peak–valley time-of-use electricity pricing factors were fully considered. Charging operations were prioritized during the specific time period of 2:00 a.m. to 4:00 a.m. each day, as this period corresponds to the deep valley of grid load, which is more favorable for improving the overall system efficiency. If the remaining available charging capacity of the energy storage system during this period is insufficient to meet the preset charging demand, the system will continue charging from other available off-peak time periods according to pre-set rules to ensure the completeness of the charging process and stable system operation. Through this refined charging strategy control and optimization, the typical daily operating load curve of the transformer and the typical daily charge–discharge curve of the energy storage system were obtained for each scheme, as shown in Figure 10 and Figure 11, respectively.

From the typical daily operating curves, it can be observed that in the ALM scheme, the charging load peak at the beginning of the grid load valley period is mainly supplied by the transformer. Therefore, the configured BESS only discharges during the grid load peak period and charges during the grid low valley period, with the operational conditions favorable for peak–valley arbitrage. In comparison, in the PLM scheme, due to transformer capacity limitations, the charging load peak at the beginning of the grid load valley period is primarily supported by the BESS. Thus, the energy storage system also discharges during the grid load valley period’s initial stage and charges in the latter part of the valley period, making the overall operating conditions less favorable for peak–valley arbitrage.

4.3. Comprehensive Economic Comparison

4.3.1. Revenue and Cost Results for the Four Schemes

This section calculates the costs and revenues for each scheme, with the results shown in Figure 12, Figure 13 and Figure 14. As shown in Figure 12, in the ALM-Smax scheme, due to transformer capacity limitations, the BESS must address both the overload issue of charging demand and the peak–valley arbitrage task. As a result, the energy storage configuration scale increases significantly, leading to the highest initial investment cost among all schemes, amounting to 8.773 million RMB. In contrast, the ALM-Smin scheme only configures the capacity for overload loads, without considering peak–valley arbitrage, resulting in a significant reduction in energy storage demand. Unlike the PLM-Smin and PLM-Smax schemes, which require large-scale energy storage investments, the initial investment cost of the ALM-Smin scheme is the lowest, at 2.584 million RMB.

Figure 13 further reveals that although the larger BESS capacity in the ALM-Smax scheme leads to higher annual O&M costs, its peak–valley arbitrage potential significantly reduces the overall electricity costs, resulting in the lowest total annual operating expense of 2.573 million RMB. In contrast, in the PLM-Smin scheme, the BESS only addresses overload issues and cannot reduce electricity costs through arbitrage, leading to the highest total annual operating expense of 3.065 million RMB.

Figure 14 shows that under all four schemes, the total revenue of the UFC station remains the same. Specifically, each scheme satisfies the same user charging demand, and both the electricity price and service fee per kWh are standardized; so, the total revenue remains consistent across schemes at 3.796 million RMB.

4.3.2. Comprehensive Economic Comparison of the Four Schemes

(1)

Step 1: Scheme Selection Based on the LCOE

The first step compares the LCOE of the four schemes to identify the most cost-effective option. Based on the method described in Section 3.2, the economic comparison results are shown in Figure 15. In terms of the LCOE, the ALM-Smin scheme performs the best, with an LCOE of 0.847 RMB/kWh, which is 5.56%, 7.22%, and 11.15% lower than that of the PLM-Smin, PLM-Smax, and ALM-Smax schemes, respectively. Therefore, the ALM-Smin scheme is recommended as the preferred option to address distribution network capacity constraints in the case UFC station.

(2)

Step 2: IRR Evaluation of the Selected Scheme

Next, the IRR of the selected ALM-Smin scheme is evaluated. According to the method in Section 3.3.2, the IRR of ALM-Smin reaches 29.2%, which is significantly higher than the commonly adopted 8% benchmark return rate in Tianjin’s charging industry. This indicates strong investment appeal and economic feasibility, fully meeting corporate investment decision criteria.

4.4. Sensitivity Analysis of the Grid Connection Cost and BESS Cost

Grid connection costs—particularly those related to transmission lines and transformers—can vary significantly across different regions and application scenarios. Meanwhile, the cost of the BESS is also subject to substantial volatility and uncertainty due to fluctuations in market supply and demand, as well as the price of key raw materials such as lithium carbonate. These two cost variables can have considerable impacts on the selection among the four schemes. Therefore, for the case UFC station, a wide-range sensitivity analysis is conducted on these two variables. The objective is to identify the decision boundaries of the four schemes within a two-dimensional cost space defined by the grid connection cost and energy capacity cost of the BESS. This analysis aims to provide a clearer reference framework for investment decisions made by UFC station developers.

4.4.1. LCOE

Figure 16 illustrates the impacts of the BESS energy capacity cost and grid connection cost on the LCOE of the case UFC station. As shown in Figure 16, under various combinations of grid connection costs and BESS energy capacity costs, the sensitivity of each scheme to these two variables varies significantly. Among the four schemes, ALM-Smax involves the largest BESS energy capacity configuration, making it the most sensitive to changes in the BESS energy capacity cost. In contrast, ALM-Smin shows minimal sensitivity to this cost due to its relatively small BESS configuration. In terms of sensitivity to the grid connection cost, since the ALM-based schemes configure larger transformer capacities, they exhibit higher sensitivity to this variable compared to the PLM-based schemes.

Based on Figure 16 above, we can further identify the optimal scheme types with the lowest LCOEs under different coordinate combinations of the two variables and generate a regional map of the most cost-effective scheme for varying cost conditions. As shown in Figure 17, the PLM-Smax scheme is more suitable for scenarios with a low BESS energy capacity cost and high grid connection cost. The critical point is when the grid connection cost exceeds 1680 RMB/kVA and the BESS energy capacity cost is below 970 RMB/kWh. The PLM-Smin scheme is more applicable to situations with a higher BESS energy capacity cost and high grid connection cost. The critical point here is when the grid connection cost exceeds 1870 RMB/kVA and the BESS energy capacity cost falls between 970 and 2560 RMB/kWh. The ALM-Smin scheme is preferred when the BESS cost is high and grid connection cost is low. The critical BESS energy capacity cost is 900 RMB/kWh, and as the BESS energy capacity cost increases, the corresponding critical grid connection cost also increases. The ALM-Smax scheme is most applicable when both grid connection cost and BESS energy capacity cost are relatively low. The critical BESS energy capacity cost is 900 RMB/kWh, and the lower the BESS energy capacity cost, the higher the corresponding critical value of the grid connection cost.

4.4.2. IRR

Figure 18 shows the impacts of the BESS energy capacity cost and grid connection cost on the LCOE of the case station.

A reference plane for the 8% IRR benchmark is drawn (in gray), with specific lines that intersect the IRR surface representing the critical values for project financial feasibility. Specifically, in the PLM-Smin scheme, the line connecting two points (3470, 0) and (530, 5000); in the PLM-Smax scheme, the line connecting two points (2510, 0) and (690, 5000); in the ALM-Smin scheme, the line connecting two points (3040, 0) and (3000, 5000); and in the ALM-Smax scheme, the line connecting two points (1860, 0) and (250, 5000) represent the critical boundaries for financial feasibility (IRR = 8%). When the coordinates of the two variables are smaller than the critical boundary, the project is financially feasible; otherwise, it is not financially feasible.

Among the four schemes, the PLM scheme has a smaller transformer capacity, and its IRR is less sensitive to fluctuations in transformer unit costs. In contrast, the ALM-Smin scheme, with a minimal energy storage configuration, is least affected by fluctuations in energy storage costs and maintains relative stability in its IRR. Conversely, the ALM-Smax scheme, which has larger configurations for both the energy storage system and transformer, is most sensitive to changes in energy storage and transformer costs, exhibiting higher volatility in its IRR.

5. Conclusions

This study systematically proposes four options and a generalized comparison framework based on ALM and the BESS to address the grid-connected power capacity constraints faced by ultra-fast charging stations, which can provide methodological references for charging operators to make decisions on specific project construction plans and for the government to formulate relevant regulatory requirements. The main contributions of this paper are as follows:

• The ALM-BESS four-quadrant classification method is proposed and four types of schemes are defined. It points out the significant differences in the transformer capacity configuration, BESS capacity configuration, and operating conditions between ALM and PLM modes when facing grid-connected capacity constraints, and finally forms four types of schemes, such as ALM-Smax, ALM-Smin, PLM-Smax, and PLM-Smin, are proposed based on the maximum and minimum configurations of the BESS. It provides a systematic and comprehensive basic scheme for supporting ultra-fast charging stations to solve the grid-connected capacity constraints.
• A generalized multi-scenario comparison process and a comprehensive economic decision-making method are proposed. Based on the given service vehicle size, charging pile configuration requirements, and grid-connected capacity constraints, a generalized selection framework is established around the four types of schemes, covering the typical daily charging load simulation, transformer and BESS capacity configurations, and typical daily working condition simulation, and a comprehensive economic evaluation based on LCOE and IRR is used to solve the problem of multi-scenario decision-making.
• The effectiveness of the proposed four-quadrant classification method and comparison framework is verified through a typical case. The effectiveness of the proposed four scenarios and the accompanying comparison framework for finding the best solution to address the capacity constraints of the distribution network is verified through a typical case of a ultra-fast charging station in Tianjin, China. The proposed multi-option comparison framework is further demonstrated to guide charging companies to make the optimal choices among the four options under different cost conditions by conducting a sensitivity analysis and solving the decision boundaries for the two key elements: the grid connection cost and the BESS cost.

From the results of the typical case study of the selected urban public ultra-charging stations, the optimal solution among the four available options will tend to be the ALM-Smax solution, as the cost of the unit grid-connected capacity decreases as well as the battery cost decreases, whereas as the cost of the unit grid-connected capacity increases and the battery cost increases, then the optimal solution among the four options will tend to be the PLM-Smin solution. While the above conclusions are in line with the trend of qualitative judgement based on common sense, this paper provides a quantitative framework for a multi-scenario comparison.

It should be noted that the multi-scenario selection framework in this paper is based on the simulation results of 96-point daily charging load profiles on a typical day, deterministic service vehicle demand, deterministic charging pile configurations at the station, the capacity configurations of the grid-connected transformer, and the capacity configurations of the BESS; additionally, the operating conditions of the transformer and the BESS for the four different scenarios are also based on static and deterministic methods, and the simulation of the user charging behaviors is also simplified to a large extent. The above simplification and approximation are mainly considered to reduce the computational complexity and simplify the solution settlement process so as to enhance the practicality of the multi-option comparison framework. However, due to the above simplification and approximation, the accuracy and generalization ability of the proposed comparison framework will be affected to a certain extent. Therefore, on the basis of the selection framework proposed in this paper, it is suggested that the next step can focus on the following two aspects to deepen the research.

• Improvement of the dynamic adaptability of the analysis: on the basis of the typical daily analysis method, the dynamic production simulation for 365 days or even years of the year will be further extended to better reflect the dynamic impacts of factors such as working days, holidays, different weather, and different seasons.
• In-depth analysis of the impact of potential risks and uncertainties: On the basis of the existing deterministic assumptions of the available capacity of the distribution network and the charging demand at different times of the day, further risk and uncertainty analysis methods are introduced, especially focusing on the impacts of uncertainties for the four schemes, such as the prediction of the available capacity of the distribution network during the valley hours, the prediction of the peak and valley arbitrage space of the energy storage, the prediction of the demand of the ultra-fast charging, the charging behaviors of users, and so on for the ALM mode. We propose the corresponding optimal configuration and selection methods.

Although there is existing space for further improvement, based on the four combinations of ALM and the BESS, this paper has formed a set of effective analytical frameworks for quantitatively comparing different combinations of ALM and the BESS through the deterministic configuration and simulation of typical daily working conditions, as well as the comprehensive economic assessment method of LCOE and IRRR, and the conclusions of this paper have good engineering practicality and interpretability, which are useful additions to the literature in this field. It also provides useful quantitative analytical methods to support the decision-making of the investment enterprises of the ultra-fast charging stations and government departments to promote the application of ALM and BESSs in ultra-fast charging stations.