Introduction
The vicinity to the earth, low-cost, relatively low complexity, and off-the-shelf elements are unique merits for LEO satellites; which have motivated researchers to consider them for communication systems. The interest in developing these satellites is increasing by leaps and bounds, mainly to be integrated with 5G terrestrial networks due to their acceptable latency values for many 5G scenarios, compared to the GEO satellites’ high latency figures. Several papers have proposed schemes that utilize the high-capacity of GEO satellites and the low latency of LEO satellites at the same time. By this we mean building communication systems that consist, for instance, of one or more GEO satellites and a constellation of LEO satellites for high data rate satellite communications [1].
Although the importance of deploying LEO satellites to support the terrestrial communication networks has been emphasized in literature [2], obtaining LEO satellites that are capable of providing the same quality of service (QoS) as that of the terrestrial network remains a big challenge. Higher-quality connections would require more complex subsystems with higher power consumption demand. The LEO satellites’ limited power resources motivated the authors of [3] to propose a power effective solution with a system comprised of GEO satellite and a constellation of LEO satellites. They suggested improving GEO satellite antennas’ receiving capabilities instead of increasing the transmitted power of the LEO satellite antennas. This, in turn, would decrease the burden of both the power consumption and the volume of the LEO satellites. It is a reasonable solution to improve the signal that reaches the LEO satellite instead of utilizing bigger or more equipment. Nevertheless, the high number of antennas in the 5G mission, which are required on the LEO satellite to support connections between the constellation’s satellites, between GEO satellites and LEO satellites, and between LEO satellites and the earth, contributes to increasing the consumed power from the Electrical Power System (EPS).
In this paper, the optimal EPS characteristics of one common type of LEO satellites (Cubesat) are investigated for the 5G mission. To this end, the CubeSat’s subsystems are defined, so the satellite’s energy requirements can be specified. Generally, CubeSats have three main segments: execution, functional, and scientific. The execution segment is related to the CubeSat propulsion and movement technologies. The scientific segment depends mainly on the mission tools and sensors, while the functional segment consists of the EPS, Communication Subsystem (COM), Attitude Determination and Control Subsystem (ADCS), and On-Board Computer (OBC) [4].
The remainder of the paper is organized as follows: Section II surveys the state-of-the-art relevant to our work and provides a list of our contributions. The on-orbit incident solar irradiance for the chosen 1U (10cm
Related Work
Previous studies have investigated the development of CubeSat subsystems and addressed different challenging design concerns. For example, the CubeSat’s high-speed movement while scanning the Earth needs effective equipment to achieve the desired altitude control [5]. In other words, an effective ADCS subsystem should alleviate the angular variations to avoid all the unexpected space disturbances [6]. The stable movement of satellites, which are orbiting based on a particular orientation scenario, allows for easier and accurate determination of the received solar irradiance to the satellite’s sides. Hence, the satellite orientation should be maintained as much as possible [7]. Other papers presented the power limitations of these satellites for the communication missions as in [8].
To determine the generated power for CubeSats, it is crucial to study the relationship between the received solar irradiance on the satellite’s sides. Hence, orientation scenarios play a key role in defining the amount of solar irradiance. In the literature, only few papers highlighted the effect of the orientation scenarios on the CubeSat performance [9], [10]. For example, the authors of [9] did not present the mathematical concepts of their work, which is focused mainly on the design of the Solar Module Integrated Converters (SMIC). In [10], the main orientation scenarios were introduced, taking into account the satellite altitude and the orbital parameters to evaluate their performance. However, the authors concentrated primarily on batteries’ characteristics with a brief discussion of the principal source of power for the satellite and for recharging its batteries, which is the satellite’s solar cells. Although [11] focused on several important aspects related to the optimal connection for the EPS components, this work did not address the solar irradiance with the orientation scenario.
Based on the fact that the incident solar irradiance excites the solar cells to generate power, it is essential to identify an accurate model of high-efficient solar panels to estimate the satellite’s generated power on orbit during all CubeSat’s trajectory stages.
The solar cells’ latest development has led to engaging materials different from silicon in the cell structure [12]. Consequently, higher conversion efficiencies were obtained by the Multi-Junction (MJ) solar cells [13]. Several materials were exploited to achieve MJ solar cells so that each material absorbs a specific range of the solar irradiance spectrum, resulting in more irradiance being converted into electrical power. A high conversion efficiency performance has been captured using MJ solar cells in [14] and [15] by mitigating the chromatic aberration. It is worth noting that Five-Junction solar cells have recorded the highest efficiencies among all PV technologies [16].
Different models were devised to define the MJ solar cell parameters in different environmental conditions [17]. In [18], the authors introduced a general review of the existing modeling approaches for the high concentration MJ solar cells in terms of the cell parameters, the received irradiance, and the temperature. On the other hand, comprehensive analytical solutions for the single-diode and two-diode equivalent circuits were employed for modeling the MJ panels [19]. While four methods for studying the MJ cell’s behavior were compared in [20], three of them are analytical, and the fourth is an algorithm for the optimization of nonlinear problems. The analytical methods in [19] and [21] presented approaches used to predict the V-I characteristics to ensure getting accurate models for the studied cells. Nevertheless, the results show some differences between the analytical and experimental validation.
Several methods of MJ parameter extraction have been investigated in the literature, such as curve fitting, fill factor, experimental results, and data-sheet information [22], [23]. In [22], the MJ parameters are obtained by curve fitting the cell’s experimental characteristics, considering the single-diode model, which leads to steady-state errors in the final values. On the other hand, the authors of [24] assumed the values of the overall series and parallel resistances, respectively, without highlighting the computation formula. In [25], the resistances were calculated depending on the empirical experiments in comparison to [26] and [27], where resistances were obtained from the V-I characteristics. Other papers did not discuss the resistances of each junction layer as in [23].
In this paper, the CubeSat power budgeting is considered as a critical enabler to the success of a CubeSat constellation for the 5G mission. A combination of related issues may affect the power budgeting. For example, it is vital to achieve a balance between a broad coverage area of the satellite network while maintaining low latency values [28], but this may influence the period of power generation. Another aspect that should be considered is the increased power consumption due to the number of used antennas [29], [30]. By this we mean, designing the EPS would experience many difficulties that should be carefully investigated. They are related to the orbit altitude, the coverage area, CubeSats high-speed movement, limited generated power under the significantly constrained size and volume resources, and 5G latency requirements [31], [32].
This paper continues the work started by the authors in [4], where several CubeSats applications were addressed besides discussing the main orientation scenarios of CubeSats with a focus on the nadir orientation scenario. In this paper, the modelling, analysis, and measurements of the EPS system are presented. By this we mean, several solar cells are simulated and the simulation results of the available solar cell are validated. Also, a new scheme for designing the EPS converters is proposed to achieve a high conversion efficiency where the previous method is employed to design and fabricate a high-efficiency boost converter. As a result, the optimal budgeting for the satellite power to meet the mission’s necessary power is reached with the presented work starting from the input (irradiance determination) to the EPS (solar cells, batteries, and converters) and ending with the output (feeding the 5G CubeSat subsystems with the needed power).
A. Paper Contributions
The main contribution in this paper, versus existing aforementioned literature, is to introduce a scheme for budgeting the power of the 5G CubeSat satellite, which can be summarized in the following points:
Employing the algorithm, presented in [4], to determine the received solar irradiance on the CubeSat’s sides for nadir orientation scenario in a simple way without considering the direct cosine matrix and Euler 321 sequences as in [10]. Taking into account the chosen height ranges that guarantee the desired latency.
Introducing an accurate model for the MJ solar cells.
Defining the type and number of batteries for the 5G mission based on the 5G mission power requirements.
Introducing a new scheme for designing high-efficiency converters.
Validating the solar cells and converter models through comparison between the simulation and the experimental results.
Designing the EPS circuit and applying the perturb & observation algorithm to obtain the maximum power point for the converters.
Optimal Satellite Orientation Scenarios
CubeSat satellites could have several orientation scenarios while they rotate around the earth, such as Nadir-Orientation, Free-Orientation, and Sun-pointing [4]. The nadir orientation scenario was chosen for the following reasons:
It is the most suitable option for the 1U CubeSat. For example, in the Sun-Pointing scenario, only one face always receives the radiation compared with the case of four faces that receive power in the nadir orientation scenario.
The 5G mission needs stable and accurate pointing towards the gateways and the User Equipment (UE) simultaneously. Hence, the free orientation scenario cannot guarantee the needed stability of connections between them.
Therefore, the nadir orientation is the potential orientation scenario for a CubeSat based 5G missions since it is always pointing towards the earth’s center, as shown in Fig. 1. In this scenario, three sides could obtain maximum power generation:
This paper uses the approach proposed in [4] to calculate the incident solar irradiance on the sides of the satellite depending on the value of the angle
The proposed angle
The aggregate generated power in (2) can be represented by:\begin{equation*} P_{in}=\sum _{s=1}^{6}P_{in_{i}}\tag{1}\end{equation*}
Since CubeSats are proceeding along the eclipse zone regularly, the power generated by the solar cells is not adequate to maintain the flow of the satellite’s operations. Moreover, in the illuminated zone, sometimes, the higher data rate demand or the increased number of users could lead to additional power consumption from the EPS. Consequently, equipping the EPS with batteries is fundamental to store enough energy for different payload conditions. Thus, the periods of illuminated and shadowed areas should be determined accurately, so batteries’ capacity could be obtained as in [4], [33].
Multi-Junction Solar Cells
MJ solar cells have an increased conversion efficiency, for a given solar irradiance, owing to the usage of multiple p-n junctions made of several semiconductors (different energy bandgaps), as illustrated in Fig. 3. Consequently, each layer absorbs a different light spectrum, which boosts the produced power. Modeling is a vital tool to estimate the generated power by the solar cells. Accordingly, the CubeSat’s generated power can be predicted before the launching process. For this reason, the following subsections will focus on modeling the Triple Junction (3J) solar cells.
Triple-junction cells structure, where
Two main equivalent circuits are used for modeling 3J solar cells. The first one is the single-diode equivalent circuit model, and it uses five parameters: the diode reverse current, series and parallel resistors, generated currents, and thermal voltage, to study the cell characteristics. The second circuit model is the three-diodes circuit, where the circuit is represented using a more complex and precise model.
A. Approximate Model With a Single-Diode Circuit
In this approach, the 3J solar cell is approximated by the one diode equivalent circuit due to the easy calculations and the reduced number of employed parameters. The single-diode equivalent circuit is illustrated in Fig. 4 [25]. Here, the cell can be described as a source, a diode, and resistors, which present the power loss in connection points and through conductors.
B. Detailed Model With the Three-Diode Circuit
This part presents the three diodes equivalent circuit for 3J solar cell modeling. As shown in Fig. 5, the equivalent circuit includes three-diode circuits connected to perform the layers of the three materials [20], [23], [24], [34]. The three circuits have a series connection that leads to the same current passing through them, and the output voltage is the sum of their voltages.
Using different materials means different saturation currents for each diode, which can be obtained using:\begin{equation*} I_{Oi}=K_{i}.\left ({\frac {T}{T_{Ref}}}\right)^{3}.\left ({\exp \left ({-\frac {q.E_{gi}}{n_{i}.K_{B}}.\left ({\frac {1}{T}-\frac {1}{T_{Ref}}}\right)}\right)}\right)\tag{2}\end{equation*}
The diode current is given as follows [23]:\begin{equation*} I_{Di}=I_{Oi}.\left ({\exp {\left ({\frac {q.V_{Di}}{n_{i}.K_{B}.T}}\right)}-1}\right)\tag{3}\end{equation*}
\begin{equation*} V_{Di}=V_{i}+I_{Out}.R_{Si}\tag{4}\end{equation*}
\begin{equation*} V_{i}=\frac {n_{i}.K_{B}.T}{q}.\ln \left [{ \frac {I_{L}-I_{Out}}{I_{Oi}}+1 }\right]-I_{Out}.R_{Si}\tag{5}\end{equation*}
On the other hand, the shunt current can be found as follows:\begin{equation*} I_{Shunti}=\frac {V_{i}+I_{Out}.R_{Si}}{R_{Sh}}\tag{6}\end{equation*}
The solar irradiance generated current is given by:\begin{equation*} I_{PV}=\frac {S}{S_{ref}}\left ({I_{L}+\frac {dI_{SC}}{dC}\left ({T-T_{ref}}\right)}\right)\tag{7}\end{equation*}
\begin{equation*} I_{L}=I_{avg}.\left ({1-A.Cos\left ({\theta _{Earth}}\right)}\right)\tag{8}\end{equation*}
The equivalent output current can be obtained as:\begin{equation*} I_{Out}=I_{Outi}=I_{PVi}-I_{Di}-I_{Shunti}\tag{9}\end{equation*}
Finally, the equivalent circuit voltage is:\begin{equation*} V_{Out}=V_{1}+V_{2}+V_{3}\tag{10}\end{equation*}
The Electrical Power System of CubeSats
In this paper, the amount of CubeSat satellites’ generated power is determined for the 5G mission. It is assumed first that a trade-off between the coverage area and the satellite altitude is realized to record low latency values. It is known that the latency requirements for the 5G connection should not exceed 1 ms for the ultra-Reliable Low Latency Communication (uRLLC) services and 4 ms for the enhanced Mobile Broadband (eMBB) services. Consequently, to meet these requirements, the satellite altitude should range between 160 km above the equator (the latency would be around 1 ms) and 600 km above the equator (the latency would be 4 ms) [32]. However, deploying CubeSats at these altitudes is not practical for reasons concerned with the atmospheric drag and reduction in the life cycle of the satellite. Thus, CubeSats can be used in applications other than time-sensitive services such as the massive Machine Type Communications (mMTC) and 5G over satellite-like broadcast and multicast.
A. The EPS Design
The knowledge of the payload energy requirements is a principle factor before designing the EPS to guarantee a successful CubeSat mission. The solar cells, batteries, and converters (current/voltage adjusting circuits) are the backbone of the satellite’s EPS. The only source of energy generation in satellites is the solar cells, whereas the storage system (backup batteries) supports the solar cells when the satellite in the shadowed zone or when the demanded energy is more than the generated one.
Several types of batteries are used for satellite missions, such as Lithium-Ion (Li-Ion), Nickel Cadmium (NiCd), and Nickel Metal Hybrid (Ni MH) [4]. Each type has its unique specifications, but the preferred type, for CubeSats missions, is the Li-Ion because it meets the constraints of mass and volume [35].
The required energy of the satellite’s subsystems \begin{equation*} E_{tot}=\frac {E}{\eta }\tag{11}\end{equation*}
\begin{equation*} E_{sun}=\int _{0}^{T_{Sun-light}}{P_{in}(t)}.dt\tag{12}\end{equation*}
\begin{equation*} E_{batt}^{r}\geq \eta .\left({E}_{sun}-\frac {T_{Sun-light}~}{T_{orbit}}E_{tot}\right)\tag{13}\end{equation*}
\begin{equation*} E_{batt}=\frac {\left ({U_{\max } + U_{\min }}\right)}{2}.C\tag{14}\end{equation*}
\begin{equation*} n=\left \lceil{ \frac {E_{batt}^{r}}{E_{batt}}}\right \rceil\tag{15}\end{equation*}
Synchronous Buck and Boost Converters
In this section, a new approach is implemented to obtain the optimal setup of the EPS converters’ parameters. The regulation of voltages/currents is achieved using synchronous boost and buck converters. Hence, the number and type of converters are defined based on the CubeSat subsystems’ voltage/current requirements. Fig. 6 displays the converters’ design process, which starts with entering the output’s desired efficiency, the frequency range, voltage/current ripples, and voltage/current margins. Then the duty cycle range is computed depending on the inserted parameters. Following that, the maximum values of inductance
A. The Converter Power Losses
The following power loss formulas were used to ensure obtaining the desired efficiency of converters. Based on the above-mentioned scheme, the power loss calculations are implemented continuously until reaching the converter’s parameters, which guarantee minimum power losses.
1) Conduction Power Losses
The power dissipation of the converter components, such as inductor and capacitor resistances, leads to power losses, which can be expressed as follows [39]:\begin{equation*} P_{cond}=P_{q1}+P_{q2}+P_{wr}+P_{esr}\tag{16}\end{equation*}
\begin{align*} P_{q1}=&{I_{eff(max)}}^{2}.R_{DS}.D \tag{17}\\ P_{q2}=&{I_{eff(max)}}^{2}.R_{DS}.(1-D)\tag{18}\end{align*}
2) Switching Power Losses
Each switching transition produces a power loss. Therefore, the total switching power loss is related to the switching frequency, and it can be obtained using:\begin{equation*} P_{sw}=P_{turn}+P_{over}+P_{qrr}+P_{gate}\tag{19}\end{equation*}
\begin{equation*} P_{turn}=\frac {2}{3}.\left ({CossL+CossH}\right).V_{min}^{2}.f\tag{20}\end{equation*}
\begin{equation*} P_{over}=0.5.I_{eff(max)}.t_{f}.f\tag{21}\end{equation*}
The power loss due to the reverse recovery current can be expressed as:\begin{equation*} P_{qrr}=Q_{rr}.V_{o}.f\tag{22}\end{equation*}
Additional power loss could be caused by charging the gate-source capacitance \begin{equation*} P_{gate}=Q_{g}.V_{max}.f\tag{23}\end{equation*}
It is worth mentioning that, the number of converters in the EPS subsystem can be reduced since it is not possible to illuminate two opposite sides of the satellite at the same time, for example,
Numerical Results
A. The EPS Design
The proposed CubeSat mission is to provide 5G services for the areas covered by its course, which means that the payload would mainly be the communication subsystem. To accomplish the mission requirements, the CubeSat should have three antennas; the first is for the feeder link, the second is an off-set parabolic reflector antenna for users link, and the third antenna would be used to relay data among the satellites of the constellation. CubeSat’s subsystems’ power and energy consumptions are displayed in Table 1 [36].
As can be seen in Table 1, the needed power is
![Right-click on figure or equation for MathML and additional features. Right-click on figure for MathML and additional features.](/assets/img/icon.support.gif)
Depending on the nadir orientation scenario, only a single side would enjoy the maximum generated power, as the solar irradiation ultimately faces this satellite’s side. Consequently, there is insignificant sun radiation received on the other sides that could be neglected. Studying the Albedo and the earth irradiance effect on the CubeSat is beyond this paper’s scope, particularly for the 5G mission. CubeSats would have low altitudes to fulfill the requirements of the 5G applications. As a result, the proceeding factors would not have significant implications on the satellite’s mission.
The net generated power, for a single side, is given by [4]:\begin{equation*} P_{in1}=2.88~W\end{equation*}
As it is evident in Table 1, both OBC and COM work almost during all the orbit period, which is logical since the payload is the communication subsystem. Assuming a 90% converters’ efficiency and from (11), the total energy needed by the satellite is:\begin{equation*} E_{tot}=19.995~W.h\end{equation*}
For a CubeSat at an altitude of 400 km, the illuminated stage period is 56.36 minutes whereas the shadowed stage period equals 36.04 minutes. As a result, the sun-light stage generated energy is computed as per (12):\begin{equation*} E_{sun}=2.705~W.h\end{equation*}
So, the energy required from Li-Ion batteries to compensate for the CubeSat needs, i.e. that which solar cells cannot cover, is:\begin{equation*} E_{batt}^{r}=17.29~W.h\end{equation*}
The characteristics of the selected Li-Ion 18650 battery are:
Nominal Voltage:
3.7~V Typical Capacity
2600~mAh The voltage’s marginal values
and4.2~V 3~V
![Right-click on figure or equation for MathML and additional features. Right-click on figure for MathML and additional features.](/assets/img/icon.support.gif)
Thus, the number of desired batteries is given as:\begin{equation*} n=\left \lceil{ \frac {17.29}{9.36}}\right \rceil =\left \lceil{ 1.84}\right \rceil =2\end{equation*}
Typically, satellites are equipped with battery packs that involve two or four batteries, such as Clyde Space and GOM Space products.
B. The Incident Solar Irradiance
After choosing the nadir orientation scenario, the generated power for each satellite side could be determined. Since the CubeSat is in a continuous movement around the earth, the gained irradiance decreases and increases based on the difference between each side angle and the solar irradiance direction [4]. Fig. 7 illustrates the variations in the CubeSat’s generated power depending on the side’s inclination angle
C. Model Verification
In a pursuit to evaluate the proposed model, five triple-junction solar cells were employed, which are manufactured by AzurSpace (3G30A), Clyde Space (CS102834), SpectroLab (XTJ), SolAero (ZTJ-
Fig. 8 shows the V-I characteristics for different 3J solar cells, which are AzurSpace (3G30A), SolAero (ZTJ-
V-I characteristics for the chosen cells at 28°C. Colored circles represent data-sheet values.
Fig. 9 illustrates V-P characteristics for the chosen cells under the STC conditions. Both the V-I and V-P characteristics demonstrate a high matching percentage between the simulated results and the nominal values of 3J solar cells that, in turn, validates the accuracy of the proposed model.
V-P characteristics of the chosen cells at 28° C. Colored circles represent the data-sheet values.
D. Verification of the Proposed Model Using Cell’s Measurements
The experiment was conducted on one of the body cells (CS102834) of the Qatar University (QU) CubeSat, produced by Clyde Space. Fig. 10 demonstrates the experiment instruments, which are:
Hydrogen light bulb.
The 3J cells.
Pyranometer sensor (Apogee SP-110), to measure the artificial irradiance on the panel.
Three lists of resistances as the load.
Multimeters.
The CubeSat side panels characteristics based-on the experimental and model characteristics.
E. The Converter Verification
The scheme introduced previously for designing converters is utilized for designing a boost converter. Table 5 compares the achieved efficiency values for several manufactured converters and the converter proposed in this paper [39], [40]. The purpose of this comparison is to prove that the introduced scheme could produce viable and practical solutions. Fig. 12 shows the synchronous boost converter circuit, which uses the Perturb and Observation algorithm
Conclusion
In this work, the basics of designing the EPS of CubeSats were addressed for the 5G mission. The power input of these satellites was covered by selecting the altitude, that considers the 5G latency, and the convenient orientation scenario, which meets the transmission consistency. Then the total amount of the received solar irradiance on the satellite’s sides was determined. Next, a highly accurate model for several MJ panels, that harvest the solar irradiance, was introduced in an effort to predict the satellite’s generated power, taking into account the efficiency of the used converters, which supply power to the satellite subsystems to guarantee the success of the mission. In addition, the solar irradiance of the satellite throughout its route, based on the nadir orientation scenario, was exhibited. Moreover, the 3J solar cells model was verified after doing measurements on the cells of QU CubeSat. Besides, the synchronous converter design was discussed and compared with other CubeSats’ converters. To conclude, this paper addressed the input, the components of the EPS, and the output of the circuit that meets the latency and power requirements of a CubeSat, designed to join a constellation for a 5G mission.
The introduced work could be enhanced further by addressing the effect of the CubeSat’s location concerning the sun and the earth in-depth, the former of which impacts the on-orbit power generation whereas the latter is concerned with the coverage area and the 5G load’s power. Using more complex algorithms to track the maximum power-point would improve the generated power amount. Finally, addressing the link budget and discussing the subsystems’ power requirements of the 5G payload would shed light on the obstructions of considering these satellites for the 5G missions and beyond.
ACKNOWLEDGMENT
This publication was made possible in part by the sponsorship agreement in support of research and collaboration by Ooredoo, Doha, Qatar. The statements made herein are solely the responsibility of the authors.