Performance Evaluation of Hybrid Battery–Supercapacitor-Based Energy Storage Systems for Urban-Driven Electric Vehicles

Abstract

Boosting the performance of energy management systems (EMSs) of electric vehicles (EVs) helps encourage their mass adoption by addressing range anxiety concerns. Acknowledging the higher power densities of supercapacitors (SCs) compared to those of the Lithium-ion (Li-ion) batteries used in EVs, this work proposes an optimal sizing and energy management strategy of a hierarchical hybrid energy storage system (H-HESS). In this system, the SCs are voltage-controlled to solely provide the current requirements of an EV motor during urban driving cycles with frequent accelerations and decelerations, while the EV battery recharges the depleted SCs. The proposed H-HESS is modeled and simulated on MATLAB/Simulink, and its performance is compared to that of a traditional battery-only energy storage system (BESS). Simulation results reveal that this H-HESS system offers a $55.7\%$ peak current reduction and ≈+2% improvement in battery loss of capacity in comparison with BESS. A pulsed battery discharge current profile is imposed by the proposed H-HESS, while C-rate control is implemented. This improves the battery aging by reducing the formation of the solid electrolytic film (SEI) that otherwise decreases its capacity.

1. Introduction

The prevalence of electric vehicles (EVs) is of a great significance to help reduce greenhouse gas emissions and reduce the dependencies on fossil fuels that are expected to become scarce in the next few decades [1]. Nevertheless, among the major factors obstructing mass adoption of EVs are the driving range limitations and the concerns on EV battery durability and lifetime [2,3]. The driving range of EVs is limited by the maximum capacity of their energy storage units. However, increasing the driving range by enlarging the capacity of the battery bank translates into higher EV prices, acknowledging that the battery size and capacity are the highest contributors to the cost of EVs [4]. On the other hand, dynamic wireless EV charging (DWC) systems enable battery capacity reduction through offering on-the-move charging [5,6,7]. However, they suffer from inherent misalignments that decrease the power transfer efficiency and cause fluctuations in the received energy, thereby degrading the performance and lifetime of the EV battery [8,9,10]. Accordingly, alternative solutions are investigated to improve the efficiency of the EV EMS, improve its battery lifetime, and reduce the EV operational costs by decreasing the frequency of battery replacement [11]. Among these solutions is the integration of supercapacitors (SC) to form a hybrid energy storage system (HESS), and developing efficient battery–SC-based EMSs to regulate the power flow between the two energy storage units and the EV motor.

SCs are electrolytic capacitors that are characterized by large capacitance values and higher power densities compared to the Li-ion batteries typically used in EVs. The higher power density allows SCs to release significantly higher power in shorter time intervals. This makes them suitable for addressing the EV motor requirements during urban driving patterns with frequent accelerations and decelerations that cause large fluctuations in the power demand [12]. At times when the current demand by the load is high, this current can be supplied by the SC to reduce the thermal stress on the EV battery and prolong its lifetime [13]. This improves the EV EMS performance and reduces its operational costs by reducing the degradation rate of the EV battery, and hence, reducing the frequency of battery replacement. The internal resistance of an SC is also significantly smaller than that of a Li-ion battery of similar energy storage capacity, which translates into lower power losses [14].

1.1. Related Works

1.1.1. Battery–SC Connectivity

Different SC connectivity solutions are proposed in the literature, to increase the operational efficiency and improve the voltage supply capabilities of the EV energy storage system [15,16]. These are demonstrated in Figure 1. One battery–SC connection configuration known as passive paralleling is proposed in Ref. [17] as shown in Figure 1a. In this method, the hybridization is formed without any power electronic circuits, which causes the SC to deplete significantly faster than the Li-ion battery. To overcome this issue, a DC/DC converter is used to connect the SC to the battery in Refs. [18,19] in a semi-active topology, as shown in Figure 1b, to control the voltage of the SC using the duty cycle of the DC/DC converter. However, in this topology, the battery bank is directly connected to the DC bus and its voltage cannot be controlled. This is addressed in Ref. [20] by swapping the positions of the energy sources in the configuration discussed in Ref. [18]. Semi-active topologies are also utilized in Refs. [21,22] where only the battery is connected to the DC/DC converter while the SC is connected directly to the DC link. In this way, the voltage of the battery can be maintained lower or higher than the SC voltage. The work in Ref. [23] compares the two connection topologies discussed in Refs. [18,20] to decide whether the battery or the SC should be connected to the DC bus through the DC/DC converters. In their assessment, the authors reveal that connecting the SC to the DC link through a DC/DC converter is better to enable the SC to supply larger current peaks.

The authors in Ref. [24] assert that the HESS should have both sources connected to the DC bus through DC/DC converters, to enable better control of their energy flow. This is demonstrated in Figure 1c. However, this suffers from higher power losses as two DC/DC converters are utilized, which decreases the overall energy exchange efficiency. In Ref. [25], the authors compare the passive paralleling topology in Ref. [17] with two improved topologies, namely, a three-level converter and a half-controlled converter. However, the two topologies utilize extra switches, and hence, require extra gating signals which increases their complexity, while the battery remains directly connected to the DC bus. Ref. [26] proposes a configuration in which the low-voltage side is held by the battery pack and is interfaced by a power diode with the DC link held by the supercapacitor bank, as shown in Figure 1d. With this topology, the battery bank is expected to operate with a lower voltage than the SC to keep the diode reverse biased. However, as the SC discharges, its voltage decreases and the diode becomes forward biased, which splits the output battery current between the motor and the SC for recharging.

Figure 1. Some topologies of battery–SC connection in the literature: (a) passive paralleling, (b) using a single DC/DC converter, (c) using two DC/DC converters (Active HESS), (d) low/high voltage isolation. © 2012 IEEE. Reprinted, with permission, from [26].

Figure 1. Some topologies of battery–SC connection in the literature: (a) passive paralleling, (b) using a single DC/DC converter, (c) using two DC/DC converters (Active HESS), (d) low/high voltage isolation. © 2012 IEEE. Reprinted, with permission, from [26].

1.1.2. HESS Energy Management Systems

Extensive research has been conducted to develop efficient EMSs for integrated battery–SC HESSs in EVs. The deterministic rule-based method is proposed in Refs. [27,28,29] in the form of if-else paradigms that are developed based on heuristic human experience. In these works, the authors set a threshold on the current demanded by the EV motor. Below the threshold, the EV battery is in operation, whereas the SC is activated when the current demand exceeds the threshold to supply the needed power. Other rule-based HESS energy management algorithms are proposed in Refs. [30,31], in which the threshold is set on the power delivered to the motor instead of its current demand. In this way, both the voltage and the current requirements of the motor are acknowledged.

In the aforementioned rule-based algorithms, the the current and/or power thresholds are determined in advance, using the expected power demand for pre-known driving cycles. This limitation is addressed in Ref. [32] by proposing an adaptive power split strategy that tracks the motor load profile in real-time to determine the level of variation in the power demand and split the power supply between the battery and the SC accordingly. Another real-time power split approach is utilized in Ref. [33], using fuzzy logic controllers to detect the load variation frequency. A fuzzy rule-based HESS EMS is presented in Ref. [34] with optimized fuzzy membership function to accurately determine the power thresholds. In addition, real-time current sensing is adopted in Ref. [35] to control the maximum current supplied by the battery to the EV motor and prevent excessive battery discharge.

Different optimization techniques can also be integrated with a rule-based algorithm to enable efficient power splitting in HESSs. In Refs. [36,37], the search space for power provisioning between the battery and the SC is first restricted by a set of rules, within which the optimal operating point is determined in real-time using metaheuristic optimization techniques, such as particle swarm optimization (PSO). Other real-time optimization solutions are also proposed in Refs. [16,38] using model predictive control (MPC) to determine the optimal power split between the EV battery and the SC bank under unknown driving cycles. Adaptive Pontryagin’s minimum principle (PMP) is also utilized in Refs. [39,40] for efficient and less computationally expensive optimization of the power provisioning in HESS. In contrast, offline, global optimal EMSs are proposed in Refs. [19,41,42] using dynamic programming (DP), yet they require prior knowledge of the driving cycle and a large computational capability of the EMS controller.

In Ref. [21], a multi-objective optimization problem is formulated to determine the optimal sizing of the battery and the SC banks required to implement a rule-based energy management algorithm in a semi-active HESS, while minimizing the cost of the HESS and extending the EV driving range. Ref. [22] also presents an optimal battery–SC sizing model for a semi-active HESS, but uses the power requirements of different driving cycles, instead of predefined EMS rules, to determine the model constraints.

1.2. This Work

In the aforementioned battery–SC connection topologies and EMS strategies, both the EV battery and the SCs are used to supply energy to the EV motor in different driving cycles, and the EMS is responsible for regulating the power provisioning between the two sources. Nonetheless, as the power and energy requirements of EVs increase to address range anxiety concerns, larger energy storage units are required, corresponding to added weight and cost of EVs. However, according to Ref. [43], with the higher power densities of SCs compared to Li-ion batteries, each additional kW required can be supplied by a supercapacitor bank that is 10–20 times lighter than a Li-ion battery counterpart. This means that, with optimal sizing of the SC bank, SCs can be utilized as primary sources of power to the EV motor during urban driving patterns with frequent accelerations and decelerations in which higher power bursts are required, without a significant addition to the total EV weight and energy consumption.

This work proposes a rule-based EMS with an optimal SC sizing model to enable an SC bank to supply the EV motor requirements during urban driving, while restricting the role of the EV battery to recharging the depleted SCs. This transforms the traditional battery-only ESS (BESS) into a hierarchical HESS (H-HESS) with voltage-controlled SCs and a current-regulated battery. The proposed rule-based EMS provides a computationally efficient approach to prove the potential of the proposed HESS with an SC as the primary power source during urban drives. The urban New York City Cycle (NYCC) is utilized to assess the performance of the EV battery and SC pack in the proposed H-HESS energy management model. The cycle aging of the EV battery is evaluated for the proposed H-HESS model and is compared to the expected aging in a BESS. In addition, the percentage reduction in the peak current requirements and the loss of capacity of the EV battery pack is compared between the proposed H-HESS and BESS in an extended 30-day simulation.

The rest of this paper is organized as follows. Section 2 presents the details of the research methodology in terms of the proposed H-HESS topology and rule-based energy management strategy, the optimal SC sizing model, and the performance evaluation metrics used in this work. A detailed evaluation of the proposed model is then presented using the simulation results reported in Section 3 and discussed in Section 4. The paper is then concluded in Section 5.

2. Materials and Methods

2.1. Rule-Based Energy Management of the Hierarchical HESS

In urban low-speed driving cycles, a recurring pattern of frequent stopping and acceleration is expected, and the maximum speed is not very high. Frequent stops and accelerations result in a high motor torque required to gain inertia, which increases the motor current, as it is typical for motors to require high starting currents. In contrast, low current is required by the motor during EV motion at a constant low speed. Nonetheless, since constant low speed travel occurs for short durations during urban driving, the SC is proposed to fully supply the motor’s energy requirements. This helps prolong the EV battery lifetime by limiting its role to recharging the depleted SCs. To fully address the motor’s requirements during urban driving, two SC banks are utilized, namely SC1 and SC2, and are connected to the EV battery and motor through gate-controlled switches, as shown in Figure 2.

A current-controlled, bi-directional DC/DC converter connects the Li-ion battery to the SC banks and another DC/DC converter connects the SCs to the EV motor. In this topology, no direct connection is established between the battery and the motor. This forms the energy supply hierarchy that allows the SC banks to supply the motor’s requirements while using the battery to recharge the depleted SCs as required. The gate-controlled switches regulate the charging and discharging of the SC banks through SC voltage control, while a current-controlled DC/DC converter ensures that the current drawn from the battery to recharge the depleted SCs does not exceed its C-rate. On the other hand, the voltage-controlled DC/DC converter connected to the EV motor ensures that the voltage across the motor is proportional to the voltage needed to achieve the desired speed profile of the driving cycle. In addition, this voltage-controlled converter helps in recharging the SCs through regenerative braking, but this is neglected in this work.

2.1.1. Voltage-Controlled SC Pack

Since the SC charge is stored electrostatically on porous electrodes, the terminal voltage of the SC is proportional to its state-of-charge (SoC), and can be used to determine whether or not an SC recharge is required, thereby controlling the operation of switches $s_{11},s_{12},s_{21}$, and $s_{22}$ in Figure 2. This is demonstrated in the flow chart shown in Figure 3.

At the beginning of the driving cycle, assuming the battery and the two SC banks are both fully charged, SC1 supplies power to the EV motor by turning on switch $s_{21}$ and keeping all other switches off, as shown in Figure 4a. During its power supply duration, the terminal voltage across the SC decreases as it supplies energy to the EV motor. As SC1 depletes its stored energy, its terminal voltage, $v_{SC1}$, significantly drops. A minimum voltage threshold, $v_{th}$, is selected to indicate the depletion of the SC bank. This is chosen to be $10\%$ of the rated SC voltage to prevent its excessive depletion beyond that point. As $v_{SC1}$ falls below $v_{th}$, the EMS turns off switch $s_{12}$ by sending a ’0’ signal to its gate, and turns $s_{11}$ to the ON state using a ‘+1’ gate signal. This enables recharging SC1 from the Li-ion battery. On the other hand, $s_{22}$ is turned to the ON state to allow the second SC bank, SC2, to supply the motor requirements, as shown in Figure 4b. As the voltage across SC2 falls below $v_{th}$, SC1 is reactivated by switching on $s_{12}$, while SC2 is allowed to recharge by switching $s_{21}$ to the ON state and switching $s_{22}$ to the OFF state, as shown in Figure 4c. The terminal voltages of both SC banks are continuously monitored by the EMS to regulate their charging/discharging processes through the gate controlled switches.

2.1.2. C-Rate Control of Li-ion Battery

As shown in Figure 2, a bidirectional DC/DC converter is connected between the Li-ion battery and the two SC banks. Pulse width modulation (PWM) is utilized to control the active switches of this converter to avoid over-discharging of the battery during the SC recharge process. This is illustrated in Figure 5. In this work, over-discharging of a Li-ion battery is assumed to take place when its discharge current exceeds twice its rated capacity, i.e., if $I_{bat}\ge 2C$. A current sensor detects the current drawn by the depleted SC, $I_{SC1}$ or $I_{SC2}$, as it is connected to the battery through its respective switch, $s_{11}$ or $s_{21}$, and compares this current value to the current threshold value, $I_{th}$, of $2C$. A PI controller is utilized to regulate the difference between the $2C$ current limit and the actual battery current, $I_{bat}$. This difference is then fed into the PWM modulator to provide the required switching signals to the MOSFET switches. This accordingly protects the battery from excessive discharge that would impact its aging performance.

2.2. Optimal Supercapacitor Sizing

An excessive increase in the EV weight by the addition of the SCs causes an undesirable increase in the EV energy consumption rates, which counters the advantages offered by the integration of SCs. To avoid this issue, the size of the supercapacitor bank must be optimized [21,44]. In the following analysis, the two SC banks in Figure 2 are collectively considered as one pack of SCs whose specifications are optimized.

Similar to battery banks, SCs can be arranged in series and parallel configurations to optimize the total resistance, total output voltage, and current. In this work, a non-linear, multi-objective optimization problem is designed with two objectives: (1) to maximize the energy that can be stored in the SCs, and (2) to minimize the size of the SC pack that can be integrated within the current mass and volume allocations for the ESS of a typical EV. This is formulated as

$$\mathrm{Maximize}{f}_1=0.5{\left(v_{{}_{\mathrm{SC}}}{n}_s\right)}^2{c}_{{}_{\mathrm{SC}}}\frac{n_p}{n_s},$$

(1)

and

$$\mathrm{Minimize}{f}_2=n_s{n}_p,$$

(2)

where $c_{{}_{\mathrm{SC}}}$ is the capacitance of each SC cell and $v_{{}_{\mathrm{SC}}}$ is its corresponding rated voltage assuming identical SC cells are used to construct the SC pack. $n_p$ and $n_s$ are the numbers of parallel- and series-connected SC cells, respectively. Multiplying $n_p$ and $n_s$ defines the total number of SC cells required in the SC pack. The first objective function, $f_1$, maximizes the stored energy in the SC bank, while the second function, $f_2$, minimizes the total number of SC cells to minimize the added weight. The following constraints are defined for the optimization problem.

$$c_{{}_{\mathrm{SC}}}\frac{n_p}{n_s}\le C_{max},$$

(3)

$$m_{{}_{\mathrm{SC}}}{n}_s{n}_p\le \gamma m_B,$$

(4)

$$V_{{}_{\mathrm{SC}}}{n}_s\le \lambda V_B,$$

(5)

$$1\le n_s\le 100,$$

(6)

$$1\le n_s\le \infty ,$$

(7)

$$n_p,n_s\in \mathbb{Z}.$$

(8)

where $m_{{}_{\mathrm{SC}}}$ and $v_{{}_{\mathrm{SC}}}$ are the mass and volume of each SC cell, respectively, $\gamma$ is the percentage of the total battery mass to be replaced by SCs, $\lambda$ is the percentage of the replaceable battery volume, and $m_B$ and $V_B$ are the total the mass and volume of the EV battery, respectively. A limit on the total capacitance of the SC bank is added in Equation (3) as $C_{max}$. By solving the MINLP model in Equations (1) and (2) subject to the constraints in Equations (3)–(8), the optimal size of the SC pack is obtained.

2.3. Performance Metrics

The use of supercapacitors to fully address the EV motor requirements during urban driving patterns is expected to enhance the EV battery performance and lifetime. This is due to the reduced stress on the EV battery by reducing its current supply requirements. To assess this performance enhancement, the current–rate-based cycle aging model in Ref. [45] is adopted, as it can detect the variations in battery aging over short driving cycles. According to Refs. [45,46], the cycle aging of a Li-ion battery depends on the moved charge, q, across its terminals during a discharging cycle. This can be evaluated as the integral of the amplitude of the battery current, $i_{batt}\left(t\right)$

$$q=\frac{1}{T}{\int }_0^T\left|i_{batt}\left(\tau \right)\right|d\tau ,$$

(9)

where T is the duration of a full discharge cycle of the battery. This is recommended to be the duration required to discharge the battery to a $20\%$ SoC. The moved charge calculated in Equation (9) is used to evaluate the decline in the state-of-health (SoH) of the Li-ion battery over the cycle duration. The SoH is an indication of the level of degradation and the remaining capacity of a battery. Essentially, it is the difference between the health of a new battery and the health of a used battery, and is expressed as a percentage or a proportion of the initial battery capacity. An empirical model for the SoH is developed in Ref. [45] by fitting a number of experimental data points for the moved charge of Li-ion batteries discharged at different discharging rates. This model is adopted in this work, as follows:

$$SoH\left(q\right)=1-5\times {10}^{-4}\sqrt{q}+2.5\times {10}^{-6}q-1.4\times {10}^{-10}{q}^2,$$

(10)

where q is the amount of moved charge calculated in (9).

In addition to cycle aging assessment, a calendar aging model is proposed in Refs. [47,48,49] to evaluate the performance of a battery over long periods of continued operation, based on variations in its temperature and SoC. This work assumes that an efficient thermal management system (TMS) is utilized to regulate the EV battery temperature. This can be implemented using microchannel heatsinks, as recommended in Ref. [50], or thermoelectric cooling as recommended in Ref. [51]. Furthermore, since this work particularly focuses on urban-driven EVs where the duration of the driving cycle is relatively short, the current-dependent cycle aging model is considered to be more suitable for performance comparison of the proposed system with a traditional BESS. The cycle aging model in [45] is further extended to evaluate the loss in capacity of the EV battery over a large number of discharge cycles, following the model in Ref. [52]. This gives a more reliable evaluation of the aging improvement offered by the proposed H-HESS. The percentage loss in capacity, ${\mathcal{C}}_{loss\%}$, is calculated using

$${\mathcal{C}}_{loss\%}=100\times \frac{{\mathcal{C}}_{ini}-{\mathcal{C}}_f}{{\mathcal{C}}_{ini}},$$

(11)

where ${\mathcal{C}}_{ini}$ is the maximum capacity of a new Li-ion battery and ${\mathcal{C}}_f$ is the final battery capacity at the end of the testing cycles.

3. Results

To simulate the performance of the proposed rule-based energy management of the H-HESS, the topology in Figure 2 is simulated on MATLAB/Simulink [53] with the proposed rule-based SC voltage control and feedback battery current control, as demonstrated in Figure 3 and Figure 5, respectively. The SC sizing problem is first solved using the metaheuristic genetic algorithm (GA) from the MATLAB Optimization Toolbox [54], then the proposed EMS is executed accordingly. The model performance is then evaluated using the performance assessment metrics detailed in the previous section.

3.1. Battery–SC Sizing

The MINLP model in Equations (1)–(8) is solved to determine the size of the SC pack to be integrated in the proposed H-HESS. The EV battery pack specifications are shown in

Table 1. Battery bank specifications.

Parameter	Value
Rated voltage	330 V
Rated capacity	$56.5$ Ah
Response time	20 s
Weight, $m_B$	625 kg
Volume, $V_B$	$0.4$ m${}^3$

, while the parameters of the optimization model are summarized in

Table 2. Parameters of the optimization model.

Parameter	Symbol	Value
Mass of each SC cell	$m_{{}_{\mathrm{SC}}}$	$0.05$ kg
Volume of each SC cell	$V_{{}_{\mathrm{SC}}}$	$42\times {10}^{-6}$ m${}^3$
SC cell voltage	$v_{{}_{\mathrm{SC}}}$	$2.7$ V
SC cell capacitance	$C_{{}_{\mathrm{SC}}}$	500 F
Percentage of total mass	$\gamma$	$7\%$
Percentage of total volume	$\lambda$	$7\%$

. The percentages of EV mass and volume, $\gamma$ and $\lambda$, are selected to be $7\%$, indicating that only $7\%$ of the weight and volume of the battery of a typical EV are replaced by the SC pack. In this way, the added SCs do not impose additional weight or space restrictions.

By executing the GA on the SC sizing model, a Pareto front of the different combinations of parallel and series connected SC cells is obtained, as shown in Figure 6. A point on the Pareto front that corresponds to $n_s=100$ and $n_p=4$ is selected because an even number of parallel SCs is required to simplify the implementation of the SC pack. The specifications of the resulting capacitor bank are summarized in

Table 3. Optimal supercapacitor bank specifications.

Parameter	Symbol	Value
Number of series SC	$n_s$	100
Number of parallel SC	$n_p$	4
Total Voltage of SC pack	$V_{S{C}_i}=v_{sc}{n}_s$	270 V
Total capacitance of SC bank	$c_{s}c\frac{n_p}{n_s}$	20 F

.

3.2. EMS Implementation

An urban driving cycle is employed to simulate the proposed H-HESS that mainly relies on the SC pack to address the motor requirements. The 600 s NYCC in Ref. [55] is considered with a maximum speed of around 33 km/h. This is shown in Figure 7.

The driving cycle in Figure 7 is applied to the motor of the H-HESS proposed in this work. The current requirements of the EV motor are plotted in Figure 8, and are observed to be quite rigorous with high peaks that reach above two the batteries’ C-rate and several rising and falling edges reflecting accelerations and decelerations. Based on the rule-based EMS in Figure 3, the operation of the voltage-controlled SCs is demonstrated in Figure 9, following the switching arrangement explained in Figure 4. SC1 is initially operated to supply the motor’s current requirements, and as it depletes, SC2 begins to operate after around 270 s while SC1 is recharged from the battery. Once SC1 is recharged, it re-operates at $t=550$ s.

The current drawn from the EV battery in the proposed H-HESS to recharge the depleted SCs is plotted in Figure 10. This is in comparison to the current drawn from an identical battery in a traditional BESS, in which the battery is solely responsible for supplying the motor’s requirements.

3.3. Aging Performance Assessment

The battery discharge current profiles in Figure 10 are used to estimate the moved charge across the battery during a full discharge cycle. A Li-ion battery with the parameters shown in Table 1 is simulated under these current discharge profile until their SoC is depleted to $20\%$. The amount of moved charge is then calculated using Equation (9) by integrating the corresponding discharge curves over their respective duration. The SoH is then evaluated using Equation (10) and the results are summarized in

Table 4. State of health of the Li-ion battery for each topology.

Topology	SoH
H-HESS	$0.9980$
BESS	$0.9962$

. As observed, the reduction in the SoH of the Li-ion is only minor due to the short discharge cycle duration, and the improvement offered by the proposed H-HESS is also minor. Accordingly, the discharge cycle of each model is repeated for a period of 30 days, equivalent to 2,592,000 s, where the EV is assumed to replicate the driving pattern until an SoC of $20\%$ and is allowed to recharge to $99\%$ then executes the same driving cycle. The loss in capacity of the Li-ion battery in each model is assessed using the simulation setup in Ref. [52]. The results are shown in Figure 11.

4. Discussion

As observed in Figure 10, the Li-ion battery in the proposed H-HESS operates only for short durations (≈27% of the driving cycle duration) to recharge the depleted SC bank. Otherwise, the battery is idle. This is in contrast to the continued operation of the Li-ion battery in a BESS. The pulsed battery discharge pattern of the proposed hierarchical HESS topology is advantageous in extending the EV battery lifetime. This is because, according to Ref. [56], pulsed discharging helps in controlling the thickness of the solid electrolyte interphase (SEI) film formed on the electrodes and limiting the overpotential of the battery, which is the difference between its terminal voltage and the internal equilibrium potential. The SEI film consumes the energy storage capacity of the Li-ion battery by causing a loss in the cyclable Lithium, thereby decreasing the battery lifetime. The pulsed discharge pattern allows the Li-ion battery to reach its equilibrium relaxation state faster and hence provides better overall battery performance. In addition, the authors in Ref. [57] reported a better utilization of the active material in Li-ion batteries charged/discharged with pulsed current profiles, which resulted in an increased battery lifetime by reducing SEI formation and surface cracking. This confirms the superiority of the proposed hierarchical battery–SC energy storage system.

Furthermore, the pulsed operation together with the C-rate control in the proposed H-HESS provide a significant improvement in the peak current requirement from the EV battery. In the results shown in Figure 10, the peak current required in a BESS is around 210 A, while the H-HESS demands only a maximum of 93 A from the battery. In addition, by observing the differences in the loss in capacity in Figure 11, the maximum battery capacity in the proposed H-HESS is only reduced by $0.85\%$, from $58.4$ Ah to $57.9$ Ah over the simulated 30-day cycle. For the BESS, on the other hand, the percentage loss in capacity is calculated to be $-2.91\%$, which may have resorted from the rigorous current profile required to address the several acceleration and deceleration phases of the driving cycle. This ≈+2% improvement in battery loss of capacity confirms the advantages of the proposed rule-based H-HESS in improving battery aging while meeting the EV driving requirements in urban driving cycles. The performance of the proposed H-HESS is also compared to a few of the other HESS topologies applied to urban driving cycles in the literature. A summary of this comparison is shown in

Table 5. Performance comparison of proposed H-HESS to existing literature (All values are calculated in reference to BESS performance).

Ref.	Driving Cycle	HESS Topology	Peak Current Reduction	${\mathcal{C}}_{\mathbf{loss}}$ Improvement
Ref. [30]	George Square cycle	Dual active	$49\%$	(not provided)
Ref. [35]	EC40 cycle	Semi active	(not provided)	≈+2% in one year (365 Days)
This work	NYCC	Hierarchical	$55.7\%$	≈+2% in 30 Days

, which highlights the performance enhancements offered by the proposed H-HESS.

5. Conclusions

This work proposes a hierarhical EV energy management system that is optimized for urban driving patterns to prevent overloading and fast aging of the EV battery. In the proposed model, SCs are utilized to fully address the requirements of the EV motor while the EV battery is used to recharge the depleted SC banks. An efficient rule-based SC voltage control algorithm is implemented to perform the switching between two SC banks to supply current to the motor. A C-rate controller is also utilized to limit the peak current demand from the battery during SC recharging. Optimal sizing of the SC pack is performed to minimize their added weight and volume. A low-speed driving cycle is used to test the proposed H-HESS using Simulink, and the results are compared to those of a BESS with no supercapacitors. A significant peak current reduction of $55.7\%$ is observed with a $2\%$ improvement in battery loss in capacity when the system is simulated for 30 days of charging and discharging. These performance enhancements come in addition to the key advantage of the proposed topology, by allowing the EV battery to only supply short pulses of low-peak current which significantly improves its aging.