Energy Harvesting Devices for Extending the Lifespan of Lithium-Polymer Batteries: Insights for Electric Vehicles

Abstract

This study rigorously evaluated the integration of energy-harvesting systems within electric vehicles to prolong battery service life. A laboratory-scale system was configured utilizing a scale electric vehicle with a 12.6 V lithium-polymer (Li-Po) battery alongside an automated control platform to precisely estimate the real-time State of Charge (SoC) through monitoring of current, voltage, and temperature of the vehicle battery under three distinct driving conditions: (A) constant velocity at 30 km/h, (B) variable velocities exhibiting a sawtooth profile, and (C) random speed variations. Wind energy was harvested employing Savonius rotor microturbines, with assessments conducted on efficiency losses and drag coefficients to determine the net power yield for each operational profile, which was found to be marginally positive. Considering the energy consumption of electric vehicles based on 2017 U.S. EPA fuel economy data, the maximal recovered energy corresponded to 0.0833% of auxiliary system demand, while the minimal recovery was 0.0398%. These results substantiated the necessity for continued research into sustainable energy management frameworks for electric vehicles. They emphasized the critical importance of optimizing the incorporation of renewable energy technologies to mitigate the environmental ramifications of the transportation sector.

1. Introduction

The widespread adoption of battery technology, while fundamental to global efforts to reduce operational carbon emissions, introduces a parallel environmental challenge. Specifically, the battery’s entire life cycle—from the initial extraction of raw materials through manufacturing to final disposal—is inherently associated with significant polluting outcomes. This study proposes a methodological approach to mitigate this trade-off by attempting to extend the functional lifespan of used batteries. The core objective is to integrate wind energy harvesters, specifically those utilizing a Savonius rotor microturbine design, as supplemental power sources. The performance of any battery system is multifaceted, necessitating the evaluation of several key parameters, among which the State of Charge (SoC) holds paramount importance. For this empirical study, the research focuses on Lithium-Polymer (Li-Po) batteries, selected for their widespread use in electric vehicles (EVs) owing to their high energy density, low self-discharge, and immunity to the memory effect. It is crucial to recognize that ambient energy harvesting devices—which convert environmental sources such as kinetic motion, thermal gradients, light, and radio frequency (RF) waves into electrical power—are not designed to replace conventional batteries. Rather, in low-power applications, they serve a vital supplementary function that reduces stress on the primary power source. Acknowledging the inherent limitations of these devices, namely their constrained conversion efficiency and lower power output compared to chemical batteries, this study explores their capacity to enhance battery lifespan. A preliminary comparison detailing the attributes of Li-Po batteries versus typical energy harvesting systems is presented in

Table 1. Comparative Features of Energy Harvesters and Li-Po Batteries.

Feature	Energy Harvester	Li-Po Batteries
Output power	Extremely low (μW to mW).	Substantially high (tens to hundreds of watts) with even higher capacities in electric mobility applications.
Energy source	Ambient, intermittent, and unpredictable (wind, solar, vibration, RF, heat).	Internal stored self-contained chemical energy.
Energy density	Extremely low.	High (150–250 Wh/kg).
Conversion efficiency	Limited (10–50%, source-dependent).	High (>90%).
Longevity	Practically unlimited (absent chemical degradation). It depends on power source availability and is affected by component wear and environmental conditions. 1	Limited (300–500 cycles typical for Li-Po).
Applications	Ultra-low power wireless sensor networks, implantable biomedical devices, and wearable wearable-mounted electronics.	Smartphones, laptops, and electric vehicles.
Maintenance	Minimal, generally self-sustaining once deployed.	Requires periodic monitoring and proper charging.
Environmental impact	Low.	Moderate to high (mining, disposal issues).

Note ¹: They experience a gradual degradation in efficiency due to aging of the materials, but they tend to maintain their functionality longer than electrochemical storage systems.

.

Table 1 presents an idealized comparison; however, device longevity varies significantly. Energy harvesters, when meticulously engineered, can sustain operational functionality for up to two decades with minimal maintenance, contingent upon the intrinsic resilience of their constituent components. Conversely, Li-Po batteries undergo progressive capacity degradation influenced by cumulative charge-discharge cycling, elevated charge-discharge rates, profound depth of discharge events, and thermal stress. Although energy harvesters exhibit superior longevity, their current power output constraints prevent standalone support of high-demand energy systems. Therefore, they are predominantly deployed within hybrid configurations designed to supplement electrochemical storage devices—such as batteries or supercapacitors—by enabling auxiliary recharging, which mitigates maintenance requirements and prolongs effective battery lifespan. Despite their incompatibility with high-power applications, advances in nanomaterials science and microelectronic engineering have expanded their applicability to miniaturized, low-power platforms, including wearable sensors and Internet of Things (IoT) nodes. Current research focuses on improving their efficiency and miniaturization [1,2,3,4,5].

1.1. Related Works

Recent investigations in energy harvesting focus on enhancing both efficiency and sustainability within vehicular and micromobility platforms. Hybrid suspension systems integrating hydraulic and electromagnetic modules have demonstrated significant improvements in ride comfort alongside increased energy recovery [6], while regenerative designs for heavy-duty trucks and magnetic mass-spring configurations harness vibrations for power generation [7]. Other approaches integrate harvesters into components such as shock absorbers and brake systems [8]. In the context of micromobility, piezoelectric and electromagnetic generators installed in bicycles convert motion into electrical energy to power sensors and portable devices, supporting greener transport solutions. The growing demand for sustainable energy in portable and wearable electronics has driven the development of self-charging systems that combine energy storage with solar, mechanical, thermal, or chemical harvesting methods to extend battery life and reduce the need for recharging. As the number of IoT devices continues to grow, battery-free energy harvesting is emerging as a promising solution to power low-energy sensors and wearables autonomously and sustainably.

Numerous studies have substantiated the role of energy harvesting as a viable complement to, or alternative for, conventional battery systems. Liu et al. [9] conducted a comprehensive review of advanced harvesting applications spanning smart homes, IoT, transportation, and aerospace sectors. De Oliveira et al. [10] proposed an RF-based power system for low-voltage batteryless devices, while Hosseini et al. [11] and Mhatre and Shukla [12] examined regenerative systems in electric vehicles, highlighting improvements in efficiency and battery life. Gao et al. [13] explored the integration of energy harvesters and energy storage for wearable and implantable medical batteryless devices. Currently, Ali et al. [14] and Safaei et al. [15] emphasized sustainable IoT powered by piezoelectric harvesters. Collectively, these studies illustrate that energy harvesting could reduce maintenance, improve reliability, and enable autonomous operation in a wide range of applications.

Several recent studies have particularly explored the integration of energy harvesting systems into electric vehicles (EVs) to boost sustainability and battery performance. Hussain et al. [16] focused on using vertical-axis wind turbines (VAWT) in EV front grilles to capture wind energy during motion and charge the battery. Broader approaches, such as that of Surendran et al. [17], have combined smart energy harvesting with automation in EVs, highlighting technologies such as wind turbines, solar panels, and regenerative braking to reduce reliance on grid charging and enhance overall efficiency. Helera and Stoichescu [18] examined the feasibility and aerodynamic optimization of vehicle-mounted wind turbines for dynamic charging to extend the EV range.

Recent review studies have provided an in-depth analysis of energy harvesting technologies for electric vehicles, including wind, solar, regenerative braking, piezoelectric, and thermoelectric systems. Hosseini et al. [11] examined the role of these technologies in battery electric and fuel cell hybrid vehicles, identifying key research gaps and future directions. Similarly, Mhatre et al. [12] evaluated the strengths, limitations, and applicability of these technologies in various scenarios. Both studies have highlighted the potential of these technologies to reduce energy consumption, extend battery life, and minimize environmental impact while addressing challenges such as low power output, conversion inefficiencies, and integration complexity. These studies also emphasized the importance of hybrid approaches that combine multiple methods, as well as future trends such as smart control systems and advanced materials to improve energy conversion and storage. In essence, these studies identified energy harvesting as a crucial element in the design of sustainable electric vehicles, albeit primarily as a supplementary rather than primary energy source.

Moreover, Savonius rotor-based wind turbines have been extensively studied due to their simplicity, low manufacturing cost, and reliable performance at low wind speeds, making them suitable for vehicle-mounted applications. Research by Dinh et al. [19] has focused on optimizing the rotor blade geometry, curvature, and auxiliary blade configurations to improve efficiency and energy output. Savonius rotors have been unable to generate enough power for propulsion, though. Arbiyani and Lasut et al. [20] reported that they were able to supply energy for auxiliary systems and marginally extend battery life. Design improvements, such as multi-curve blades and aerodynamic enhancements, have significantly increased performance compared to traditional designs. These findings highlight the potential of Savonius rotors as practical energy harvesters for EVs, particularly in urban environments with variable wind conditions.

1.2. Research Context

No research indicates the total substitution of batteries with energy harvesters. Research underscores the incorporation of energy harvesters into hybrid systems to augment energy supply and prolong battery longevity, especially in low-power devices. These systems help manage intermittent power sources and allow for smaller battery sizes. Current efforts primarily explore how energy harvesters can complement batteries. The key strategies are summarized below:

• Research and development to improve energy harvesting efficiency. Innovations such as triboelectric nanogenerators and hybrid systems improve power output by utilizing multiple energy sources. For instance, Khan et al. [21] developed a solar optimizer based on maximum power point tracking (MPPT), while Liao et al. [22] explored cellulose-based harvesters. Mondal et al. [23] and Sevcik et al. [24] reviewed hybrid devices and energy transducer technologies. Further research focuses on the advancement of thermal and solar harvesting techniques [25,26,27,28].
• Hybrid systems for IoT and wearable devices. Various studies investigate the integration of energy harvesting in IoT systems and wearable devices to reduce the need for battery replacement and optimize energy use [29,30]. Abdulmalek et al. [31] developed a hybrid wearable healthcare system for real-time monitoring. Nishanth and Senthilkumar [3] proposed using solar or piezoelectric harvesters to extend the battery life of the fitness tracker. Ali et al. [32] reviewed wearable energy harvesting methods, focusing on the heat and motion of the human body, and examined various technologies, including piezoelectric, triboelectric, thermoelectric, and hybrid systems.
• Electric and hybrid vehicles. In recent years, several studies have examined how best to utilize the energy of electric and hybrid vehicles. Particularly, Prasad et al. [33] introduced a method that integrates an optimization algorithm with a neural architecture to determine the optimal energy distribution. In addition, Manivannan [34] applied machine learning to create a smart energy management system for hybrid electric vehicles. To enhance its performance, an IoT-based smart charging system was implemented to schedule vehicle-to-grid connections. Finally, Shen [35] studied the trends in IoT-based charging stations for electric vehicles and their roles in smart energy dispatch, load balancing, remote monitoring, and the integration of renewable energy. Furthermore, the study highlighted the primary challenges of scaling and implementing IoT-based systems, including cost ineffectiveness, interoperability issues, and security concerns.
• Advanced energy storage and harvesting systems. Portable energy storage and harvesting systems are vital for daily use, especially in healthcare and wearable technology. Traditional batteries tend to be bulky, but advances in materials have allowed for flexible, lightweight alternatives. These integrated systems support continuous operation and reduce dependency on external power. Zhang et al. [36] reviewed technologies such as solar cells, biofuel cells, nanogenerators, super-capacitors, and various batteries, analyzing energy density, power, and durability. To address the mismatch between intermittent energy harvesting and constant power needs, advanced storage solutions such as super-capacitors are being explored [37,38,39,40,41]. Their fast charge/discharge capabilities and long cycle life make them a promising option for hybrid systems [42,43,44,45,46,47,48,49].
• Smart energy management systems. Hybrid systems are based on smart circuits to efficiently manage the energy of both harvesters and batteries [50,51,52]. These systems dynamically balance power sources to ensure a stable supply and maximize the use of harvested energy [53,54,55,56]. In this sense, Yaseen et al. [57] developed a framework for resilient IoT systems using dynamic energy management and sustainable harvesting methods.

1.3. Contribution

Among the five focus areas outlined in Section 1.2, this study centers on “Electric and hybrid vehicles.” To that end, a case study was performed using a scaled electric vehicle prototype, three representative driving profiles, and an integrated control system. The primary contributions of this study are as follows.

• Experimental data link driving profiles to the battery SoC in an electric vehicle. It supports safe and efficient operation, helps estimate battery life through charge-discharge cycles, and enables discharge curve analysis to assess the available energy.
• Test findings on the application of wind energy harvesters as supplementary power sources in small EVs are presented. These results help estimate the potential improvement in the SoC and lifespan of the battery.
• Reckoning how much of the total energy consumed by an electric vehicle can be attributed to energy harvesters based on the 2017 U.S. EPA fuel economy data [58,59].
• A detailed assessment is conducted to quantify efficiency losses and estimate the net energy contribution of the energy harvesters. This assessment is based on the technical specifications of the Savonius-rotor microturbine.

This study specifically addresses the integration of renewable energy within electric vehicles, concentrating on energy harvesting through Savonius-type wind microturbines and real-time battery SoC monitoring. The research scope encompasses the development of sustainable energy management strategies for EVs, with particular attention to the aerodynamic implications of the energy harvesting devices and the auxiliary power demands of the EV systems. Additionally, this study adds value by examining the integration of energy-harvesting devices into EVs, an innovative approach to improving sustainability and energy efficiency. It further contributes by quantitatively assessing feasibility through real-time SoC measurements across different driving profiles and by calculating net power after accounting for drag and efficiency losses. Therefore, this study intends to provide empirical evidence rather than theoretical assumptions. It also identifies technical limitations such as low energy recovery and aerodynamic drag of energy harvesters and sets the stage for further improvements in the design and integration of these devices.

This study is systematically organized to fulfill its research objectives. Section 2 delineates the theoretical framework, details the employed methodology, describes the case study, and specifies the experimental conditions. Section 3 presents the results derived from three distinct driving profiles, elucidating the contribution of energy harvesters to battery energy demands. Additionally, this section discusses the strengths of the experimental design, alongside the study’s limitations and potential future directions. Section 4 provides a comparative analysis of energy consumption, benchmarking wind-based harvesting against RF and solar energy harvesting systems. Finally, Section 5 offers a comprehensive conclusion, synthesizing the principal findings of the investigation.

2. Materials and Methods

2.1. Considerations

A battery converts chemical energy into electricity using cells composed of an anode, cathode, and electrolyte. Its theoretical capacity depends on the amount of reactive material. As one of the costliest components of an electric vehicle, monitoring the SoC is crucial for estimating remaining energy, travel range, and battery health. SoC management helps prevent premature degradation, and analyzing the discharge curve reveals how voltage drops under different driving conditions.

This study focuses on the battery discharge curve, a key tool for evaluating the performance, efficiency, and lifespan of a battery under three different driving conditions. Engineers use it to design battery management systems (BMS) that handle voltage variations and determine cut-off voltage to prevent over-discharge. As the battery ages, shifts in the curve indicate degradation, providing insight into when a replacement will be needed. Manufacturers provide discharge curves to guide battery selection.

2.2. Method Description

A ten-step method was implemented to rigorously assess the proposed energy harvesting. This method was applied to three distinct EV driving profiles established on the docking platform. The primary goal was twofold: to estimate the battery SoC and to determine the requisite number of energy harvesting circuits needed to balance power consumption and generation. SoC indicates the remaining battery capacity but not power demand. From a battery perspective, it is essential to accurately estimate the SoC, as it is used to prevent the battery from overcharging or over-discharging, optimize its lifespan, and estimate its State of Health (SoH). Operationally, accurate SoC estimation for EVs enables the reliable prediction of vehicle autonomy, facilitating efficient route planning and stop logistics.

Balasingam et al. [60] categorized SoC estimation into three different approaches: (i) current-based, (ii) voltage-based, and (iii) combined methods using nonlinear filters, such as the extended Kalman filter. For the present study, the Coulomb counting method was chosen due to its simplicity and cost effectiveness. This current-based technique requires only a sensor and integration logic, enabling real-time monitoring during charge/discharge without resting periods. It tracks charge flow over time and applies to various battery types [61,62].

The proposed method is outlined as follows:

• Fully charge the vehicle battery to its nominal capacity.
• Connect the electric vehicle to the docking platform, ensuring a secure coupling.
• Select a driving profile and start the electric vehicle to begin the experimental procedure.
• Acquire the voltage, current, and temperature of the vehicle battery.
• Save the data to the microSD memory card, enabling subsequent review via the organic LED screen integrated within the electronic system.
• Estimate the battery SoC using the Coulomb counting method as defined by Equation (1)

  $$SoC\left(t\right)=So{C}_0-{\displaystyle \frac{1}{C_{nom}}}{\int }_{t_0}^{t}I\left(t\right)\mathrm{d}t,$$

  (1)

  where $So{C}_0$ represents the initial battery SoC at full capacity (100%), $I\left(t\right)$ is the instantaneous battery current, and $C_{nom}$ denotes the nominal battery capacity.
• Terminate the experiment once the battery voltage, $V_{bat}$, approaches the predefined cut-off voltage, $V_{co}$.
• Compute the instantaneous power consumption as $P\left(t\right)=V\left(t\right)I\left(t\right)$.
• Ascertain the power output ($P_{harv}$) of a single energy harvester using the manufacturer specifications or design-based calculations, accounting for the efficiencies of both the power source and conversion circuitry.
• Determine the minimum number of energy harvesting circuits ($N_{min}$) required to sustain continuous system operation, ensuring that the aggregated power output meets or exceeds the system’s average energy demand, formally expressed as:

  $$N{P}_{hav}\ge P_{cons}.$$

  (2)
  Note that if $P_{harv}$ = 0, then N = 0, since there is no energy available for harvesting. Therefore, $N_{min}$ should be estimated according to Equation (3) when $P_{harv}$ > 0.

  $$N_{min}=\lceil P_{cons}/P_{harv}\rceil ,$$

  (3)

  where $\lceil ·\rceil$ indicates that the number should be rounded up to the next whole number, as using only a fraction of an energy harvesting circuit is not feasible.

This method provides three key outcomes: (i) battery SoC, (ii) system power consumption, and (iii) the number of energy harvesting circuits required. In this approach, the battery acts as a buffer between the intermittent power supply and the system’s energy demand. SoC is the primary metric for validating the accuracy of the energy supply system design.

2.3. Case Study

This section presents a case study in which the voltage and current supplied by a lithium-ion battery were measured to monitor and control the systems of a small-scale electric vehicle. The total energy consumption was quantified to estimate the battery discharge curves. The battery was installed in the 1:10 scale model of the electric vehicle shown in Figure 1a, whose technical specifications are shown in

Table 2. Technical specifications of the electric vehicle, battery, and motor used in the case study.

Component	Parameter	Specifications
Electric vehicle	Length	379 mm
	Width	200 mm
	Height	62 mm
	Weight	1.44 kg
	Wheel diameter	66 mm
	Transmission	Single speed direct drive
Battery-Traxxas 3 Cell	Technology	Li-Po
	Total capacity, $C_{nom}$	5000 mAh
	Nominal voltage	11.1 V
	Continuous discharge	125 A
	Maximum explosion speed	250 A
	Watt Hours	55.5
	Loading rate	5 A
	Maximum load rate	10 A
	Dimensions	155 mm × 26 mm × 44 mm
	Weight	376 g
	Discharge cut-off voltage 1	9.25 V
	Temperature range (charging condition)	0 °C to 43 °C
	Temperature range (discharging condition)	0 °C to 60 °C
	Temperature range (storage condition) 2	15 °C to 25 °C
Velineon 3500 motor	Technology	Brushless
	RPM/volt	3500
	Max RPM	50,000
	Magnet type	Neodymium
	Connection type	3.5 mm
	Current	65 A to 100 A
	Diameter	36 mm
	Length	55 mm
	Weight	262 g

Note ¹: It refers to the battery voltage when its $SoC\left(t\right)$ approaches 0. %Note ²: It is the condition of a cool, dry place that does not expose the battery to direct sunlight, and whose temperature does not exceed 60 °C.

.

In this case, “Small-scale” refers to the reduced-size models used for testing. Note that this electric vehicle consists of mechanical, electrical, and electronic components. It was coupled to the docking platform shown in Figure 1b with bearings that allow the vehicle’s wheels to rotate without moving the vehicle.

Figure 2 presents a schematic diagram of the docking platform used to perform the experiments. This platform was utilized with the 1:10 scale model of the electric vehicle shown in Figure 1a. In this diagram, all subsystems of the docking platform (as seen in Figure 1b) are labeled. Furthermore,

Table 3. Specifications of the three sensors used.

Sensor	Parameter	Technical Specifications	Additional Information
ESP32 ADC	Voltage	Range: 0–33 V	Voltage divider 1:10
		Sensitivity: 3.3 mV/LSB
		Resolution: 10 bits
Modified INA219	Current	Range: 0–32 A	Modified for high currents
		$R_{Shunt}$: 0.01 Ω (original 0.1 Ω)	Accuracy: ±1%.
		Resolution: 15 bits
KY-013	Temperature	Range: −55 °C to 125 °C	Thermistor NTC 10 kΩ
		Accuracy: ±0.5 °C.	Linearized by software
		Low energy consumption	Resolution: 10 bits
		Dimensions: 22 mm × 15 mm × 9 mm	Steinhart-Hart coefficients:
			A = 0.001129148,
			B = 0.000234125, and
			C = 0.0000000876741

offers specifications for the sensors utilized in the experiments, which are detailed in the box labeled “Electronic system for measuring battery voltage, current, and temperature” of Figure 2.

The calibration of the modified INA219 sensor was performed using software by calculating two essential parameters in the device: $Current\_LSB$ and $Cal\_Value$. Firstly, $Current\_LSB$ is the smallest current that can be measured, which determines the measuring range, and it should be estimated using a practical approximation defined by Equation (4)

$$Current\_LSB={\displaystyle \frac{I_{max}}{2^{15}}},$$

(4)

where $I_{max}$ represents the maximum current that should be measured in the experiments; in this case, $I_{max}$ = 32 A, and 2¹⁵ represents the range of the current registered on the chip since its resolution is 15 bits.

$Cal\_Value$, the integer written to the calibration register, can then be calculated using Equation (5). This equation accounts for the maximum differential shunt voltage (in Volts) that the modified INA219 can handle, which is 0.04096 V, an internal fixed reference that ensures proper scaling. The operator $\lfloor ·\rfloor$ denotes the integer part function, and $R_{Shunt}$ refers to the low value resistor used to prevent instrument damage and facilitate the measurement of higher currents. It is the value of the external shunt used to develop the differential voltage across the input pins.

$$Cal\_Value=⌊{\displaystyle \frac{0.04086}{Current\_LSB-R_{Shunt}}}⌋.$$

(5)

The KY-013 sensor uses an NTC (Negative Temperature Coefficient) thermistor, which decreases resistance as temperature increases. Its analogous nature and sensitivity can lead to fluctuations that appear as temperature changes, causing inaccurate readings. These inaccuracies are influenced by the resolution of the analog-to-digital conversion and the linearization by software. Since the response of the thermistor is nonlinear, the Steinhart-Hart equation should be used to linearize and calibrate the sensor by determining its specific coefficients, as shown in Equation (6). Note that this equation represents a model relating the varying resistance, $R_T$, of a device to its varying temperature, T.

$$T={\displaystyle \frac{1}{A+Bln{R}_T+C{\left(ln{R}_T\right)}^3}},$$

(6)

where

$$R_T=R_f\left({\displaystyle \frac{V_{cc}}{V_{out}}}-1\right).$$

(7)

$R_f$ is the reference resistor, $V_{CC}$ is the supply voltage, $V_{out}$ is the voltage measured at the analog pin, and A, B, and C are the Steinhart-Hart coefficients indicated in Table 3.

Although ideal calibration involves measuring resistance at three known temperatures using a high-precision thermometer, this case used the recommended values from the sensor datasheet, assuming a nominal resistance of 10 kΩ. The typical reference points are ice water (~0 °C), warm water (~40 °C), and room temperature (~25 °C) to calculate the Steinhart-Hart coefficients. The electric vehicle uses a Velineon 3500 brushless motor with Neodymium magnets and a coil-based stator, controlled by an electronic drive that regulates speed and power. A radio transceiver manages control signals during docking tests. The motor drive acts as an actuator, and its shaft connects to a gear system that increases torque, enabling bidirectional wheel movement.

2.4. Experimental Conditions and Settings

This section describes the fundamentals and key components of experimental design. The fundamentals are the assumptions that ensure the validity and reliability of the experimental design, including unbiasedness, sufficiency, applicability, and replicability. Key components include defining variables, identifying experimental units, and using the procedures for measurement.

Firstly, to ensure the validity and reliability of the experimental design, three driving profiles were defined. These driving profiles were generated by a software module based on an external microcontroller, which was programmed in C to generate three digital signals that control vehicle speed. It stores three driving profiles in nonvolatile memory.

• Profile A. It is characterized by a constant speed of approximately 30 km/h.
• Profile B. It features a variable speed according to a sawtooth signal.
• Profile C. It entails a driving profile with randomly varying speeds.

Figure 3 shows the speed of an electric vehicle for three driving profiles.

Additionally, to ensure the validity and reliability of the experimental design by addressing aspects such as unbiasedness, sufficiency, applicability, and replicability, the variables were continuously monitored for 40 min, yielding 2404 samples for each at an approximate sampling frequency of 1 Hz. In the context of energy harvesting and electrochemical systems, extended single-run experiments are considered valid because these systems often exhibit both rapid transient responses and slower steady-state dynamics that must be captured simultaneously. A sufficiently long monitoring time ensures that transient phenomena, such as start-up instabilities or short-term oscillations, are recorded. Meanwhile, the sampling resolution contributes to ensuring that steady-state behavior and recurring cycles are adequately characterized. Therefore, this strategy captures both the fast transient responses and the repetitive steady-state cycles inherent to conduction profiles, thereby providing a comprehensive representation of system behavior [63,64].

This methodological choice aligns with reliability engineering practices, which commonly employ prolonged single-run experiments to evaluate degradation trajectories and performance consistency in battery systems [65]. In this study, repeatability was confirmed for driving profile B; during the 40-min monitoring period, the phenomenon repeated approximately every 3.1 min, resulting in 13 reproducible cycles within the same run due to the sawtooth signal. Thus, in addition to the statistical robustness provided by the sample size, the internal repetition of the phenomenon provides further evidence of reproducibility and enhances the dataset’s validity. Taken together, these methodological and analytical considerations demonstrate that the single, long-duration experiment provides statistically representative and scientifically credible evidence to support the reported results.

To strengthen replicability, two electronic systems were employed: one for measuring the current, voltage, and temperature of the vehicle battery, and the other for recording vehicle speed. The digital voltage generated for each driving profile was subsequently converted to an analog format using a digital-to-analog converter and transmitted to the electric vehicle via an RF transceiver.

For the electronic system measuring battery parameters, it was assumed that the electric vehicle operates with a 12.6 V, 5000 mAh Li-Po battery. Accordingly, battery voltage, current, and temperature were monitored through sensors during docking tests, with data displayed on an OLED screen and stored on a microSD card for subsequent analysis. All tests were conducted within safe temperature ranges from 0 to 43 °C for charging and from 0 to 60 °C to ensure reliable data on power consumption, as detailed in Table 2.

For the electronic system implemented to monitor vehicle speed, a photoelectric sensor positioned behind the wheel detects light interruptions caused by notches on the wheel. The microcontroller processes these digital pulses to calculate RPM, converts the values to km/h, and displays the speed on an OLED screen. In this way, the independent variables are battery voltage, current, and temperature, as well as vehicle speed. The dependent variable is the battery SoC.

This study considered three experiments, each aligned with driving profiles A, B, and C. The Li-Po battery was fully charged to 12.6 V, with a cut-off voltage of $V_{co}$ = 9.25 V to prevent damage. The tests were carried out at ambient temperatures between 21 and 25 °C. A KY-013 sensor, linearized using the Steinhart-Hart equation, was placed on the battery surface. The ESP32 (10-bit ADC) sampled data every 5 s.

3. Results

3.1. Experiment Based on Driving Profile A

Figure 4 presents the voltage, current, temperature, and SoC of the battery for driving profile A. As shown in Figure 4a, the autonomy of the electric vehicle was 36 min, at which point the battery voltage reached $V_{bat}$ = 9.58 V—exceeding the specified cutoff voltage $V_{co}$. In this case, the vehicle speed was around 30 km/h. Additionally, Figure 4a shows that the current demand ranged from 6.5 A to 11.0 A, and the temperature increased from 21.5 °C to 26.75 °C.

Moreover, fluctuations in temperature measurements were evident, attributable both to the limitations of the KY-013 sensor, as detailed in Section 2.3, and to variations in the battery’s current demand from driving profile A. To better model the behavior of the battery current, a sigmoidal fit was used with a BiDoseResp model (R² = 0.77) to capture the trend of non-linear temperature. There are notable inflection points at approximately 25 and 31 min that show distinct, sequential heating phases, which are characteristic of this driving profile. Additionally, Figure 4b illustrates the battery SoC behavior, highlighting that the SoC remains at 1.80% while the battery voltage $V_{bat}$ stays above the cut-off voltage, $V_{co}$, thereby safeguarding against battery damage.

3.2. Experiment Based on Driving Profile B

Under driving profile B, the vehicle operated for over 39 min at a variable speed between 0 and 44 km/h, following a sawtooth signal (see Figure 3). Figure 5 displays the corresponding battery voltage, current, temperature, and SoC throughout this profile. As shown in Figure 5a, the EV autonomy lasted 39 min, at which point the battery voltage reached $V_{bat}$ = 9.38 V, slightly above $V_{co}$. During this period, the current demand fluctuated between 5.0 A and 15.0 A, while the battery temperature rose from 23.0 °C to 27.0 °C.

Similar to driving profile A, the temperature fluctuations observed in driving profile B stemmed from both the inherent limitations of the KY-013 sensor (refer to Section 2.3) and variations in the battery’s current demand. Subsequently, a polynomial regression (R² = 0.19), shown in Figure 5a with the dotted red line, was used to model the trend of nonlinear temperature, reflecting a steady increase from 22 °C to 25 °C. Note that a small coefficient of determination (R²) was obtained due to the high variability of the measured temperatures. In contrast to driving profile A, driving profile B exhibited significant temperature changes due to sensor quality issues and abrupt signal fluctuations caused by the sawtooth signal. However, the trend in the temperature signal was preserved. Therefore, this behavior in the battery temperature represented the overall thermal response to driving profile B. Note that $SoC$ = 5.11% and $V_{bat}\ge V_{co}$, which prevented damage to the battery.

3.3. Experiment Based on Driving Profile C

For driving profile C, the vehicle operated at random speeds between 0 and 48 km/h for 40 min (see Figure 3). Figure 6 presents the voltage, current, temperature, and SoC of the vehicle’s battery for driving profile C. As shown in Figure 6a, the EV autonomy was 39 min, at which point the battery voltage reached $V_{bat}$ = 9.68 V—exceeding $V_{co}$. In this case, the current demand ranged from 9.0 A to 15.0 A, and the temperature increased from 18.0 °C to 28.0 °C.

As with driving profiles A and B, temperature fluctuations in driving profile C were linked to the limitations of the KY-013 sensor (see Section 2.3) and variations in the battery’s current demand. Subsequently, a peak function model, shown in Figure 6a with the dotted red line, was used to fit the non-monotonic temperature trend (R² = 0.28), indicating partial capture of thermal fluctuations. Despite limitations, the model confirmed real transient heat transfer events, offering a useful framework for analyzing temperature behavior under dynamic conditions. Note that $SoC$ = 4.44% and $V_{bat}\ge V_{co}$, which prevented battery damage.

Finally, Figure 7 shows the distances traveled by the electric vehicle for each driving profile, and

Table 4. Comparison of the interest variables for the three driving profiles.

Parameter	Profile A	Profile B	Profile C
Final $SoC$ (%)	1.80	5.11	4.44
Time (min)	36	39	40
Final $V_{bat}$ (V)	9.58	9.38	9.68
Battery temperature (°C)	21.5 to 26.75	23 to 27	22 to 28
Traveled distance (km)	17.49	13.48	17.12

presents a summary of the results obtained.

3.4. Integration of Wind Energy Harvesters into the Li-Po Battery System

This section outlines a 1:10 scale electric vehicle powered by a Li-Po battery, supported by a 100 mm Savonius rotor-based electric wind microturbine. The turbine featured a 4-vane, 50 mm-high fan and a high-speed motor (3–12 V DC, 0–25,000 RPM, 130 magnetic brushed torque). The motor body measures 27 mm in diameter and 38 mm in length, with a shaft of 10 mm × 22 mm. As shown in Figure 8, the vertical vane is able to capture energy from any wind direction. This setup followed the procedure in Section 2.2.

Given these considerations, it was essential to determine $P_{cons}$, $P_{harv}$, the efficiency losses of the microturbine, and $N_{min}$ for each driving profile.

• Estimation of $P_{cons}$. Figure 9 shows the instantaneous power consumption of the 1:10 scale electric vehicle under each driving profile. The black, red, and blue lines represent profiles A, B, and C, respectively. Instantaneous consumed power, $P_{cons}\left(t\right)$ = $V\left(t\right)I\left(t\right)$ was calculated from the voltage and current of the battery over time.
• Determine $P_{harv}$ and the efficiency losses of the microturbine. In the first term, considering that the microturbine was based on a Savonius rotor [66]. $P_{harv}$ was given according to Equation (8).

  $$P_{harv}=P_{wind}{C}_p{\eta }_{glo},$$

  (8)

  where ${\eta }_{glo}$ = ${\eta }_{mec}$${\eta }_{gen}$, with ${\eta }_{mec}$ representing the mechanical efficiency and ${\eta }_{gen}$ denoting the generator efficiency of the microturbine.
  It is assumed that $C_p$ is the power coefficient and $P_{wind}$ is the available wind power calculated using Equation (9), where $A_T$ = $H\times D$, H is the height of the rotor in m, D is the diameter of the rotor in m, $\rho$ is the air density in kg/m³, and v is the wind speed in m/s.

  $$P_{wind}={\displaystyle \frac{1}{2}}\rho A_T{v}^3.$$

  (9)
  Assuming that ${\eta }_{mec}$ = 0.95 due to the direct gear system with minimal friction, ${\eta }_{gen}$ = 0.7 reflecting the simplicity of the microturbine electronics, and $\rho$ = 0.923 kg/m³ at 27 °C and 2000 m altitude, $P_{harv}$ was given by Equation (10). Since the power coefficient of the Savonius microturbine is usually between 0.15 and 0.25, a worst-case scenario was assumed, and $C_p$ was set to 0.15.

  $$P_{harv}=0.9975P_{wind}.$$

  (10)
  Similar to Figure 9, Figure 10 shows the harvested electric power, $P_{harv}$, for driving profiles A (black), B (red), and C (blue).
  Using battery power data (Figure 9) and microturbine output (Figure 10), Figure 11 illustrates how the integration of two microturbines reduced the power demand of the Li-Po battery over time across driving profiles A, B, and C, assuming that $P_{Diff}\left(t\right)$ = $P_{wind}\left(t\right)-2P_{harv}\left(t\right)$. Figure 11 shows that by integrating two microturbines, the energy demand was reduced to 97.21%, 98.06%, and 95.64% for the driving profiles A, B, and C, respectively. In this way, the use of two microturbines suggests that the battery life could be extended by up to 2.79%, 1.94%, and 4.36% for driving profiles A, B, and C, respectively. This gain is proportional to the reduced energy demand from harvested power.
  As demonstrated in Figure 12, the results indicate enhancements in the SoC of the battery, attributable to the energy yield of the microturbine-based energy harvester. Note that SoC’ represents the estimated SoC when the wind energy harvester was used in the EV.
  To confirm that the microturbine rotational speed, n, ranged from 0 to 25,000 RPM as indicated in Section 3.4, Equation (11) was considered.

  $$n={\displaystyle \frac{60\lambda v}{\pi D}}.$$

  (11)
  Note that $\lambda$ is a dimensionless quantity representing the specific speed of the microturbine. For a Savonius system, $\lambda \approx$ 1, since the tangential speed at the end of the rotor vanes is nearly equal to the wind speed. During the experiments, n ranged from 0 to 2800 RPM, which is well within the operating range of 25,000 RPM for the motor. The efficiency losses for any wind speed could be estimated using Equation (12).

  $$L=1-{\displaystyle \frac{P_{harv}}{P_{wind}{C}_p}}=1-{\eta }_{mec}{\eta }_{gen}=0.335.$$

  (12)
  Note that L was estimated using the worst-case scenario for a Savonius microturbine, with ${\eta }_{mec}$ = 0.95 and ${\eta }_{gen}$ = 0.7.
• Determine $N_{min}$. Figure 13 shows $N_{min}$ calculated using Equation (3), which accounts for how the speed of the electric vehicle impacted $P_{harv}$ for each driving profile.

  Table 5. Comparison of the results considering the three driving profiles.

  Parameter	Profile A	Profile B	Profile C
  Average of $P_{harv}$ [W]	1.35 ± 0.26	1.08 ± 1.30	1.48 ± 1.47
  Minimum of $N_{min}$ [devices]	0	0	0
  Average of $N_{min}$ [devices]	71	327	405
  Maximum of $N_{min}$ [devices]	122	4905	4672

  shows a summary and comparison of the results from the three driving profiles.
  Note that with only two microturbines ($N_{min}=2$) and assuming $P_{cons}=100$ W under driving profile A, each harvester must supply at least 50 W. To achieve this, the EV must maintain a speed greater than v = 27.88 m/s (100.36 km/h). However, such conditions increase power demands under other profiles.

3.5. Aerodynamic Drag Introduced by Microturbines

In EV applications, it is crucial to consider the aerodynamic drag introduced by Savonius microturbines. These turbines harvested wind energy while operating, but they also increased the aerodynamic drag coefficient of the vehicle. This increase in aerodynamic drag led to higher energy consumption and could counteract the energy that was harvested. Therefore, this study assessed the impact of utilizing Savonius microturbines by determining whether the energy savings exceeded the energy costs caused by the additional drag.

According to the prevailing view, the Savonius turbine is a drag-based machine. This means that it operates based on the difference in drag force between its two blades. Therefore, a turbine of this type, when mounted on an EV, especially at the front or top, acts as an additional obstacle, increasing the aerodynamic drag coefficient. Thus, the net energy equals the energy generated by the two turbines minus the additional energy consumed due to increased aerodynamic drag, as indicated by Equation (13).

$$P_{net}=2P_{harv}-\Delta P_{drag}.$$

(13)

Assuming that,

$$\Delta P_{drag}={\displaystyle \frac{1}{2}}\rho A_{EV}{v}^3\Delta C_d,$$

(14)

where $\Delta C_d$ is the increase in the EV drag coefficient due to the microturbine, and $A_{EV}$ is the EV frontal area.

Thus,

$$P_{net}={\displaystyle \frac{1}{2}}\rho v^3\{2A_T{C}_p{\eta }_{mec}{\eta }_{gen}-A_{EV}{\displaystyle \frac{\Delta C_d}{{\eta }_{mec}}}\}.$$

(15)

For $P_{net}$ > 0, $p_{harv}$ > $P_{drag}$, and consequently, Equation (16) must be satisfied.

$$\Delta C_d\le {\displaystyle \frac{2A_T}{A_{EV}}}{C}_p{\eta }_{mec}^2{\eta }_{gen}.$$

(16)

Thus, according to Equation (16), for the harvested energy to be relevant to the EV used in this case, a microturbine contributing with a $\Delta C_d\le$ 0.038211 must be considered.

To validate this condition, $\Delta C_d$ must be determined by considering one of three approaches: (i) experimental measurements placing the EV with the turbine in a wind tunnel, (ii) using computational fluid dynamic (CFD) to simulate airflow around the vehicle with and without the turbine, and (iii) empirical estimation when the experimental or CFD data are unavailable.

Based on specialized studies, an empirical estimation was applied in this study. El et al. [67] used the shear stress transport (SST) k-$\omega$ turbulence model in ANSYS-Fluent to simulate a vehicle with a front-mounted horizontal-axis wind turbine, recovering 5.13% of energy. The vehicle (4625 mm × 1730 mm × 1482 mm), modeled in SOLIDWORKS, showed a drag coefficient increase of 4.0512% (from 0.2814 to 0.2928) at a wind speed of up to 27 m/s (97.2 km/h) while traveling 100 km. Similarly, Yildiz and Besir [68], using the Viscous-Standard k-$\epsilon$ turbulence model, reported a 6.5% energy recovery and a drag coefficient increase of only 0.3864% (from 0.5693 to 0.5715) for a vehicle model (4000 mm × 1810 mm × 1482 mm) that included a front-mounted horizontal-axis wind turbine, as used by El et al. [67], operating at a wind speed of 27 m/s (97.2 km/h) while traveling 100 km. Finally, to determine the optimal location for the wind microturbine on the vehicle (front, hood, or roof), the study by Sofian et al. was reviewed [69]. Using ANSYS-Fluent, they modeled a sedan vehicle (4200 mm × 1700 mm × 1800 mm) to determine its drag coefficient while traveling at 25 m/s (90 km/h). They demonstrated that the drag coefficient of the vehicle changed by 0% at the front, 15.3846% at the hood, and 30.7692% at the roof. In other words, the drag coefficient increased from 0.39 to 0.45 at the hood and to 0.51 at the roof. However, the drag coefficient did not change when the turbine was located at the front of the vehicle.

Thus, following Sofian et al. [69], the microturbine was positioned at the front of the vehicle. Based on the studies by El et al. [67] and Yildiz and Dandil [68], the drag coefficient ($C_{d_0}$) ranged from 0.28 to 0.57 for a vehicle, and the change ($\Delta C_d$) ranged from 0.0039 to 0.0405. For this study, the midpoints were selected; that is, $C_{d_0}$ = 0.4250 and $\Delta C_d$ = 0.0222.

Consequently, based on Equations (13) and (14), $P_{net}$ > 0 at any EV speed, since ${\eta }_{mec}$ = 0.95, ${\eta }_{gen}$ = 0.7, and $C_p$ = 0.15 (See Equation (17)).

$$P_{net}=2.0898\times {10}^{-4}\rho v^3.$$

(17)

This result confirms the findings of previous studies, such as those conducted by Rogowski et al. [70] and Yildiz and Dandil [68], which concluded that aerodynamic drag is a significant factor in the use of microturbines in EVs. These studies also suggested that the net energy generated is marginal and may be insufficient to offset the increased aerodynamic drag.

4. Discussion

4.1. Energy Requirements in an Electric Vehicle

To determine the baseline energy needs for a scaled vehicle, this study used 2017 EPA fuel economy data, highlighting the superior efficiency of electric vehicles over gasoline models. The energy use aligns with

Table 6. Energy consumption of electric vehicles based on 2017 U.S. EPA fuel economy data [58,59].

Usage and Losses	City & Highway	City	Highway
Charging battery	10%	10%	10%
Auxiliary electric	0–4%	0–6%	0–2%
Energy to wheels	65–69%	60–66%	71–73%
Electric drive system	18%	20%	15%
Accessories	3%	4%	2%
Idle	0%	0%	0%
Energy recovered from regenerative braking	22%	34%	6%

, particularly for a 2012 Nissan Leaf, which accounts for the charging losses estimated by the Argonne National Laboratory.

Thus, according to Equation (17), the energy harvester has a marginal capacity to generate energy to cover a portion of a vehicle’s accessory energy demand in various driving scenarios, including combined city-highway, city, and highway conditions.

Table 7. Power consumed by the EV for each driving profile, net power generated by the wind energy harvester, percentage of power covered by the energy harvester, and percentage of covered energy of the required energy by auxiliary electric systems in EV.

Driving Profile	${\mathit{P}}_{\mathit{cons}}$	${\mathit{P}}_{\mathit{net}}$	${\mathit{P}}_{\mathit{cov}}(\%)$	Energy Covered (Auxiliary Electric, %)
A (average values)	92.2880	0.1102	0.1194	0.0398
A (max values)	121.6675	0.2050	0.1685	0.0561
B (average values)	66.6996	0.0888	0.1331	0.0444
B (max values)	165.9492	0.3646	0.2197	0.0732
C (average values)	79.6508	0.1208	0.1517	0.0506
C (max values)	173.3424	0.4332	0.2499	0.0833

shows the proportion of the energy demand covered by this energy harvester for each driving profile and scenario. These results demonstrate how the energy harvested provides a marginal contribution to the energy demands of electric vehicles. While this harvested energy is insufficient to eliminate reliance on batteries or enable full battery-free sustainable mobility, it partially meets the energy needs of electric vehicles. Consequently, this technology holds promise for low power applications, including the Internet of Things and wireless sensor networks. Nevertheless, current limitations in power density and efficiency underscore the need for further research into scalable hybrid solutions. Additionally, reducing dependence on lithium involves exploring alternatives such as Sodium-ion batteries and improved recycling methods. These findings show that energy harvesting systems could be used in new ways to strengthen electric mobility, supporting a more sustainable and diversified transportation future.

4.2. Comparison Between Wind Energy Harvesters and Ambient RF Energy Harvesters

An analysis comparing micro wind turbines, ambient RF, and photovoltaic energy harvesting systems is required. All of them harvest energy from the environment, but they differ in size and use.

Table 8. Wind energy harvesters versus ambient RF energy harvesters [71,72,73,74,75,76,77].

Feature	Wind Energy Harvesters (Small-Scale)	Ambient RF Energy Harvesters
Energy source	Kinetic energy of moving air.	Ambient electromagnetic energy.
Power output	mW to W. Capable of powering more demanding electronic systems.	μW to tens of μW. Suitable only for ultra-low-power devices.
Environment	Operates most effectively in outdoors.	Operates reliably both indoors and outdoors.
Physical size	Typically ranges from few cm to meters.	Generally sub-centimeter in scale.
Maintenance	Higher. Subject to wear and tear.	Lower/None. Solid-state components with no moving parts.
Appearance	Potential visual and acoustic pollution even at small scales.	Invisible and silent.
Applications	Remote sensing and monitoring, IoT, low-power street lighting.	Wireless sensor networks, RFID tags, implantable medical devices, and battery-free IoT devices.

shows the important aspects to compare, such as energy source, power output, environment, physical size, maintenance, appearance, and applications.

Wind energy harvesters offer higher power output but require consistent wind and are often large and complex, making them suitable for large-scale applications. In contrast, ambient RF harvesters offer reduced power consumption but are compact, low-maintenance, and effective in wireless environments, albeit with limitations for ultra-low-power applications. In the context of electric vehicles, wind microturbines have emerged as a more viable energy harvesting option compared to RF systems. Wireless charging systems [78], such as the Q_i standard for smartphones, are not energy harvesting systems. Consequently, a direct comparison between wireless charging systems and the energy harvesting systems described in this study is not possible. While wireless charging systems are designed for energy transfer, the systems used in this study are intended for energy harvesting.

4.3. Comparison Between Wind Energy Harvesters and Photovoltaic Harvesters

According to

Table 9. Factors for selecting wind or photovoltaic energy harvesters [79,80,81,82].

Feature	Wind Energy Harvesters (Small-Scale)	Photovoltaic Energy Harvesters
Energy source	Kinetic energy of moving air.	Solar radiation.
Time dependence	Supplies energy day and night with sufficient wind.	Output varies with light intensity and cloud cover.
Form factor and size	Requires a large swept area Miniaturization is difficult.	Well-suited for integration on surfaces like roofs or casings.
Location suitability	Areas with consistent, high-speed wind.	Areas with high solar irradiation.
Space/footprint	Uses less land, allowing space for farming or grazing.	Needs a large, flat, shadow-free area for optimal exposure.
Installation/Maintenance	Higher. Subject to wear and tear.	Lower complexity and easier installation.
Environmental impact	May cause noise, visual impact, and pose risks to birds if poorly sited.	Primarily associated with panel production and land use in large setups.
Efficiency	Can achieve high conversion rates, limited by Betz Law.	Typical module efficiency is lower, but systems are often easier to install and scale.
Scalability	Challenging—small turbines are inefficient at low wind speeds.	High scalable—small cells power low-duty sensors effectively.

, the most suitable type of energy collector—wind or photovoltaic—depends on at least nine factors. This implies that each type works best under different conditions.

Microturbine-based wind energy harvesters are typically large, despite research on smaller, non-rotating alternatives such as piezoelectric and aeroelastic systems. Conversely, because they are small, silent, and simple to incorporate into surfaces, photovoltaic energy harvesters are favored for urban use. While wind energy harvesters are more appropriate for off-vehicle recharging, photovoltaic energy harvesters are more useful for electric vehicles. The two methods are contrasted in relation to electric vehicles in

Table 10. Comparison between wind and photovoltaic energy harvesters [83,84,85,86].

Feature	Wind Energy Harvesters (Small-Scale)	Photovoltaic Energy Harvesters
Energy harvesting	Inefficient—harvested energy is minimal and outweighed by added drag.	Highly efficient—harvests stable energy while parked or driving in sunlight.
Integration	Low feasibility due to bulky, noisy turbines that add drag and reduce battery range.	Highly feasible—easily integrated into flat surfaces like the roof or hood.
Aesthetics and safety	Low. Not only is it visually disruptive, but its moving parts also pose safety and noise hazards.	Excellent. Flat, sleek, and safe.
Practical power	Very slow. Small turbine energy is negligible in electric vehicle power use.	Moderate. It can extend driving range or supply power to low-demand auxiliary systems.

.

Despite the clear advantages of photovoltaic energy harvesters for electric vehicles, this study aims to establish a baseline and define the context for exploring Venturi microturbine–based energy harvesters. The achievement of a positive net energy gain with these microturbines poses significant engineering challenges.

4.4. Advantages of Experimental Design

The experimental design used in this study offers four main advantages that strengthen its validity and applicability. First, it incorporates real-time monitoring of key battery parameters—current, voltage, and temperature—under controlled conditions, ensuring accurate estimation of the battery SoC. Second, the use of three distinct driving profiles (constant, sawtooth, and random speeds) provides a comprehensive representation of real-world driving scenarios, enhancing the generalization of the findings. Third, the integration of an automated control platform allows for precise data acquisition and minimizes human error, improving reliability and replicability. Additionally, the design accounts for aerodynamic drag and efficiency losses when evaluating energy recovery, which adds rigor by considering practical constraints rather than idealized conditions. Finally, continuous monitoring over an extended period ensures that both transient and steady-state behaviors of the electrochemical system are captured, resulting in a robust dataset for analysis.

4.5. Perspectives and Limitations

This research has the potential to stimulate innovative strategies for the sustainable integration of renewable energy into electric vehicles. Although current wind energy recovery is minimal, future improvements in turbine design, aerodynamic optimization, and lightweight materials could significantly increase the efficiency of energy harvesters. Furthermore, combining wind energy with other renewable sources, such as solar panels or regenerative braking, along with implementing intelligent energy management systems, could transform this concept into a practical solution for powering a vehicle’s auxiliary systems. These advances would reduce the environmental impact of transportation and support the development of greener mobility technologies.

However, this approach is not yet commercially viable due to the extremely low amount of captured energy, below 0.05% of the auxiliary system’s energy demand. Aerodynamic drag and efficiency losses limit its practicality. While the concept demonstrates technical feasibility, substantial research and development are needed to overcome these limitations and achieve significant energy gains.

5. Conclusions

This study employed a docking platform integrating an electric vehicle with three systems: a driving profile module, an electronic system for monitoring battery-related parameters, and an electronic system to track the speed of a scaled-down electric vehicle. Utilizing current, voltage, and temperature data from the Traxxas 3-cell Li-Po battery under three distinct driving profiles—the SoC was estimated. Despite significant differences in driving conditions, SoC behavior remained consistent across all profiles. To avoid battery damage during testing, continuous monitoring of voltage and temperature was critical. However, the high sensitivity of the temperature sensor occasionally led to inaccuracies, indicating a need for more reliable sensors in future experiments. Two Savonius microturbine-based energy harvesters were used. Although the energy harvested was insufficient to meet the total energy demand of the system, it could reasonably support the accessory electric systems of the vehicle. Notably, in the optimal scenario—using the driving profile C on a highway—the energy harvested was 2.18 times the energy required by the accessories. Using the driving profile B (city scenario), in the best case, the recovered energy could supply up to 0.0833% of the auxiliary electric system demand, and 0.0398% in the worst case, using two microturbines. This projection was based on the 2017 EPA Fuel Economy Data, considering that the energy distribution among the components of the electric vehicle was 3%, the midpoint in the declared range. Even though the recovered energy is marginal (0.0833% at best), the study emphasizes sustainable energy management strategies for EVs, aligning with global efforts to reduce transportation-related emissions. During energy harvesting, the Savonius rotor’s RPM was monitored to ensure operation within the technical specifications of the microturbine. Although considered inefficient and low in energy output, a Savonius-based microturbine was chosen because of its advantages: low rotational speed, high torque, low sensitivity to wind direction, and a simple design. With these features, this microturbine is appropriate for driving in a city. Based on these factors, a prediction was made indicating that, depending on the driving profile, the harvested energy could marginally extend the battery life. In addition, the number of microturbine-based wind energy harvesters required to fully meet the energy needs of the system was estimated, yielding averages of 71, 327, and 405 devices for driving profiles A, B, and C, respectively, without considering losses due to drag forces in EVs when the microturbine is added. A comparative analysis was also conducted between microturbine-based energy harvesters, ambient radio frequency (RF) harvesters, and photovoltaic systems. The analysis showed that each type of harvester has its own advantages and disadvantages, depending on the specific application and context. The following key factors were considered in the comparison: energy source, functionality, power output, environmental conditions, size, maintenance, and aesthetics. Finally, aerodynamic drag emerges as the principal factor limiting the efficacy of Savonius turbines (or any other type of turbine) for substantial power generation on a moving EV. Thus, reducing the vehicle’s overall drag coefficient remains the most effective strategy for extending EV range, rather than incorporating additional components that increase aerodynamic resistance.