Li and Shi EURASIP Journal on Wireless Communications and

Networking (2015) 2015:179

DOI 10.1186/s13638-015-0414-2

RESEARCH Open Access

An intelligent solar energy-harvesting

system for wireless sensor networks
Yin Li and Ronghua Shi*

Abstract
An intelligent solar energy-harvesting system for supplying a long term and stable power is proposed. The system
is comprised of a solar panel, a lithium battery, and a control circuit. Hardware, instead of software, is used for
charge management of the lithium battery, which improves the reliability and stability of the system. It prefers to
use the solar energy whenever the sunshine is sufficient, and the lithium battery is a complementary power supply
for conditions, such as overcast, rain, and night. The system adapts a maximum power point tracking (MPPT) circuit
to take full advantage of solar energy, and it ensures the lithium battery an extremely long life with an appropriate
charging method, which shortens the frequency of the battery charge-discharge cycle. This system can be
implemented with small power equipment which is especially suitable for outdoor-based wireless sensor nodes
in the Internet of Things (IOT).
Keywords: Solar energy harvesting; Bimodule power supply; MPPT; WSN; Internet of Things

1 Introduction The state-of-the-art energy-storage techniques for

Wireless sensor network (WSN) is the second largest energy-harvesting systems in sustainable wireless sensor
network after the Internet in the world, and it ranks as nodes can be classified into two technologies, i.e., super-
the first of the next ten emerging technologies. Cur- capacitors and rechargeable batteries [9]. These two cat-
rently, it has been used widely in the Internet of Things egories have their own advantages and disadvantages,
(IOT), mainly for environmental parameter monitoring involving energy-storage density, lifetime, discharging,
in various production circumstances, such as greenhouse leakage, size and so on [10]. Since the supercapacitors
[1, 2], water quality monitoring [3, 4] and so on. have significantly lower power density and higher leak-
Conventionally, disposable batteries can be used for age overhead than rechargeable batteries [11], which
power supply in WSN, where researchers have made ef- makes them impractical for small-package WSN nodes,
forts to save the finite battery on power control by routing we employ an energy-harvesting system using a lithium
algorithm and topology optimization [5–7]. On the other battery as the storage.
hand, reducing the power consumption of the nodes al- From the electrochemical theory, we may learn that
ways sacrifices performances like computing. The most the aging of the lithium battery is influenced greatly by
up-to-date power density of available battery technology the self state of charge (SOC) [12–14]. Data from [15]
cannot match the needs of most WSN for long lifetime show that lithium batteries in high SOC are more vul-
and small form factor, which limits the use of WSN due to nerable to environmental impacts of aging and SOC cyc-
the need for large batteries. It also has a slight possibility ling the batteries enhances the resistive lives. Therefore,
that the better batteries for small devices will become it must be avoided that the battery always be in high
available in the next few years. Energy harvesting and SOC to extend battery lifetime, and approaches, one ex-
management may be the most convenient ways to solve ample of which is recharging the battery until its voltage
the problem of making WSN autonomous and enable drops below a specific level, should be taken [16]. The
widespread use of these systems in many applications [8]. charging managements are usually employed by micro-
controllers for the flexibility of software designing and
* Correspondence: [email protected] implementation [9, 17], but researchers [18] have proven
School of Information Science and Engineering, Central South University, that battery-charging controlled by software may have
Changsha 410083, China

© 2015 Li and Shi. This is an Open Access article distributed under the terms of the Creative Commons Attribution License
(http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly credited.
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 2 of 12

some problems in which charging logic could not work, WSN usage, such as solar power, wind, vibration, heat,
and that the battery could not be charged under sufficient and RF [9]. How to collect and store energy effectively
sunlight. In our system, the charging management is im- from the environment has been taken more and more
plemented by hardware instead of the codes running seriously by researchers.
within the microcontroller for consideration of reliability. Sharma et al. [21] studied a sensor node with an
This paper focuses on an intelligent solar energy- energy-harvesting source, and a buffer was used to store
harvesting (ISEH) system based on maximum power the generated energy. The sensor node periodically
point tracking (MPPT) for wireless sensor nodes used in sensed a random field and generated packets. Only when
IOT, which prefers to use the solar power and takes the the energy was available, the packets would be transmit-
lithium battery as a supplementary under the condition ted; otherwise, they were stored and waited upon. They
of inadequate illumination. To prolong the lithium bat- also exploited throughput optimization, i.e., to obtain
tery life, an intelligent circuit using RS triggers is pro- energy management policies for the largest possible data
posed, which makes the lithium battery charge only rate and the minimal mean delay in the queue. In order
when the battery voltage is lower than a specific value. to increase the lifespan of WSN with powering-up
The circuit can be divided into two main functional methods, Ramasur et al. [22] took some efforts in a wind
parts, i.e., the charging sub-circuit and the control sub- energy harvester (WEH) model. Their WEH consisted of
circuit. The sub-circuits are merged into one printed cir- a wind generator and a power management unit to store
cuit board (PCB), and the whole system has been de- and condition the generated energy. The results showed
signed, built, and tested. Experimental results show that that their aero-elastic flutter generator could produce
the system can work stably and quite fit the require- more power compared with that of the other small-scale
ments of pre-designing. wind generators; however, the circuit was not equipped
The contributions of this paper for the ISEH systems with maximum power point tracking (MPPT) resulting
are as follows: in poor efficiencies of the WEH.
Besides harvesting the wind power, taking full advan-
(1)Charging control of the lithium battery is tage of the solar power may be more convenient in
implemented dexterously by RS triggers, which WSN usage. Although solar power is time- and season-
supports a reliable and stable operational status for dependent, it remains as one of the best choices by
the system contrasting the approach executed by adapting a power management mechanism [23]. Yi et al.
software. [24] put forward a wireless sensor node design based on
(2)We combine the advantages of other solar-harvesting a solar energy-power supply. They gave full consider-
systems, which are dominated by solar power using a ation to energy-saving principles by adapting low-power
lithium battery as an energy storage and only when consumption devices in every module and collected the
the battery voltage drops below a specific level before solar energy to provide lasting power for the system.
charging it. This architecture will help extend the life Another low-power solar energy-harvesting system
of the battery and the system, and avoid wasting the for WSN was put forward in [25], where Naveen et al.
solar energy. employed the system in an intelligent building. They
adapted a solar energy harvester instead of an alkaline
The construction of this paper is organized as follows. battery for the sensor nodes. They used a number of solar
In Sec. 1, we give an introduction of our study. A brief cells connected in series and parallel to each other to scav-
review on the related work is outlined in Sec. 2. Then, enge energy, and they applied a set of ultracapacitors to
we present the problems in solar energy-harvesting sys- store up the energy. As a backup energy source, alkaline
tems and the proposed system construction in Sec. 3. In batteries were connected along with the capacitors.
Sec. 4, the corresponding calculations of some important A solar-powered sensor module using low-cost capac-
parameters are shown. In Sec. 5, we provide the func- itors as storage buffers was investigated in [26], and only
tional module implementation of the system; simulation capacitors were adapted. The advantage is that the
and experimental results are also included in this sec- energy-charging time is shortened within a second, al-
tion. In Sec. 6, a conclusion of this paper is made. though the module cannot work without light illumin-
ation. A battery-supercapacitor hybrid energy-storage
2 Related work module was proposed in [17], and an embedded proces-
Self-sustainable WSN systems are on the verge of being sor was used to control the charging of the battery.
a broad requirement in many fields [19, 20], since most Supercapacitors charge the battery only when their
of WSN applications are difficult to maintain after their saved energy exceeds the peak requirements of proces-
deployments. Researchers make great efforts to find out sors running at full speed and ignores estimating the
renewable energy resources from the environment for SOC of the battery.
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 3 of 12

Taneja [27] et al. brought forward a kind of micro photo- According to the survey of the related work, many re-
voltaic energy system. Alberola et al. [28] put forward searchers provide better solutions to the first two points
another solar power system consisting of a supercapaci- mentioned above; however, the last point on how to im-
tor and a lithium battery. To provide an uninterrupted prove the stability and reliability of the circuit needs to
power supply, the literature [29] exploited two battery be researched further. Common energy management
groups for energy storage. In charging lithium batteries, and charge control are always achieved by software,
some literatures have been presented [30–32]. The com- while in cases of extreme discharges, the microcontroller
mon method is that whenever the voltage of the solar itself may be powered off and cannot be restarted even if
panel is high enough, charging starts. According to the there is sufficient illumination, leading to the question,
research of Ecker et al. [15], Takahashi et al. [33], and how could the battery be charged? In this paper, we
Liu et al. [34], the lithium battery capacity could fade focus on application of hardware realization of charge
rapidly when it is in high SOC for a long term; thereby, management, which can greatly improve the robustness
Jiang et al. [16] and Li et al. [35] have proposed a solu- of the solar-harvesting system.
tion that the system charges the lithium battery only
when the voltage is lower than a specific value. This 3.2 System construction
strategy can solve the problem of charging too often The ISEH system is physically composed of a solar
and the lithium battery being in high SOC all the time; panel, a lithium battery, and a control circuit. The con-
nevertheless, it may waste most of the solar power in trol circuit comprises a solar MPPT module, a charging
that this system applied the lithium battery as the pri- sub-circuit, an over-discharged protection sub-circuit,
mary power source no matter how capable the solar and a boost DC/DC module for the lithium battery. The
panel was to supply the system or not. system schematic diagram is shown in Fig. 1.
As shown in Fig. 1, the system has three input branches,
3 Methodology i.e., a solar panel, a lithium battery, and a mini-USB inter-
In this section, to begin with, we analyze some shortages face. It also has an output branch offered by an ordinary
of the existing solar energy-harvesting systems and then USB interface. In this paper, we use “bimodule” to
present the methodology and the hardware components emphasize that the output of the system may originate
of our system. from the solar panel or the lithium battery, respectively.
The standard mini-USB interface is a reservation to
3.1 Problems to be considered charge the lithium battery by an external power adapter if
Energy harvesting is one of the most promising tech- necessary. The functional components of the system are
nologies toward the goal of perpetual operation of demonstrated as follows:
WSN. Recent developments have allowed renewable en-
ergy sources such as solar or wind power to be used for (1)Solar power supply
wireless sensor nodes. Concerning the usage of solar en- We propose a new solution for supplying the power
ergy, a lot of scholars have conducted considerable re- to the sensor nodes, contrasting with that in [28]
search works, while there are still some aspects which and [31]. In those two papers, the task of the solar
could take more optimization. power is only for charging the lithium battery and
the super capacitor by a DC/DC converter, which
(1)Making full use of the solar energy is very may cause part of the solar energy to be wasted.
important. The system employs the solar power as While in our work, the solar branch has the priority
the preferential power source as long as the to provide electrical power to make full use of its
sunshine is available [27–29], rather than that which energy unless overcast, rainy days, or night etc.
takes the rechargeable battery as the primary one comes. The power generated via this branch flows to
and applies the solar energy only for charging. the MPPT module first. The MPPT guarantees that
(2)To extend the rechargeable battery life as far as the system can utilize the almost peak power
possible and keep the high performance of the produced by the solar panel, and the principle of
battery, the charging process of the battery MPPT are shown in Fig. 2.
should be taken into control, avoiding too many When the solar panel meets a light load or comes
charge-discharge cycles or the battery will always very close to opening a circuit, the output voltage
be in a high-charge state. may approach the highest level. The power and the
(3)Designing a simple and ingenious control circuit can current are simultaneously very small. As the load
reduce the complexity of system development, becomes heavier, there are many changes that will
decrease the power consumption, and increase the be made until it reaches the peak value and then
stability and reliability of the system. gradually declines. For example, the output voltage
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 4 of 12

Fig. 1 Scheme for the ISEH System

of the solar panel being decreased, the current being drives the switch (2) to connect to port B, which
increased more rapidly, and the power being disconnects the solar branch and connects the
increased gradually as a synthesis. The MPPT circuit lithium branch to the system output. A protection
helps the output voltage of the solar panel stay circuit is placed right after the battery to avoid
around the peak power area. Therefore, the system overdischarge, followed by a DC/DC boost circuit
can apply it with maximum efficiency [36, 37]. to raise the voltage from battery nominal voltage to
(2)Lithium battery power supply system output voltage. Most of the values used in
As shown in Fig. 1, a mechanical single-pole single-throw this paper are listed in Table 1.
(SPST) switch (3) is located following the lithium (3)Charging circuit
battery, which can deter the lithium battery from In Fig. 1, the switch (1) is an electronic SPST which
wasting energy in an open circuit and keep it safe controls the charging status of the lithium battery. It
in the case of transportation or stock. After system maintains the switch-off status as default, which
deployment, the switch should be switched on means that the battery is not charged. If and only if
manually. If the power provided by the solar panel the battery voltage is underneath the predefined
cannot satisfy the load, the voltage comparator (2) level, the voltage comparator (1) triggers the switch

Fig. 2 Power, voltage and current of the solar panel

Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 5 of 12

Table 1 Summary of values Table 2 Power consumption of one sensor node

Voltage/V Current /mA Vin/ V I/ mA P/ mW T/ h W/ Wh
Battery nominal voltage 3.7 Cortex-m0 5 4 20 2 0.04
System output voltage 5 ZigBee 5 40 200 2 0.4
Average operating current 16 Sensors 5 48 240 2 0.48
Output voltage of the solar panel 0–23 Sum 0.92
Normal voltage 5
Voltage threshold 3.6
Protection voltage 3.3 data given in the table are the maximum or close to it.
Each of the sensor’s current consumption is set to the
Danger boundary 3
average operating current. The notation T stands for the
working hours in one day, 2 h in a day means the duty
cycle of 8.33 %, that is to say, a data monitoring period
(1) on, and the battery is charged at once if the solar lasts for 5 min every hour. As a kind of statistical result,
panel could afford it. As soon as the charging starts the power consumption of one sensor node is about
for a few seconds, the battery voltage rises rapidly 0.92 Wh every day.
and changes the state of the comparator (1). That is
why a RS trigger is required to keep the charging 4.2 Battery selection
status. The control chip of the charging circuit is Comparison of common rechargeable batteries is shown
CN3063 [38]. As charging is terminated, the pin in Table 3. Numbers 1–5 represent lead-acid battery,
END of the chip changes to low level, which turns nickel-cadmium battery, nickel-hydrogen battery, lithium-
the switch (1) off and disconnects the charging cir- ion battery, and lithium-polymer battery, respectively.
cuit. The switch status is also maintained by the RS As demonstrated in Table 3, the lithium-polymer bat-
trigger. tery is relatively satisfactory for the system usage, whose
capacity depends on the capacity of the load and the
4 System parameters analysis longest rainy days of the system-deployed region.
4.1 Power consumption of node
In applications of IOT, WSN nodes are generally divided 4.3 Calculation of battery capacity
into the following three categories; the sensor nodes, the The calculation formula of the battery capacity (BC) is
routing nodes, and the sink node. The power consump- given by
tion of the sink node is usually the largest, but it can be
BC ¼ A  QL  NL  TO=ðcc  VÞWh ð1Þ
deployed indoors or easy to approach outdoors, so it is
unnecessary to consider its power supply. In terms of In this expression,
hardware construction, the sensor nodes usually have
more sensors than the routing nodes, for example, the A - Safety factor, between 1.1–1.4
dissolved oxygen sensor, PH sensor etc. The total power QL - The average daily power consumption of the load,
consumption of the sensor group is larger than that of Wh
the sensor node itself, consumed by data computing and NL - The longest continuous rainy days, set 7 according
radio transmission, so that the power consumption of to the experience
the sensor nodes is larger than that of the routing nodes. TO - Temperature correction factor, in general, TO = 1
In the system design, power consumption of the sensor when temperature is above 0 °C, and TO = 1.1 above −10 °C
nodes should be taken into the overriding consideration. and TO = 1.2 below −10 °C.
In our research, the sensor nodes usually include the cc - The depth of battery discharge, generally speaking,
following components: if it is a lead-acid battery, the value is 0.75, if it is a
nickel-cadmium battery, the value is 0.85, and if it is a
(1)ARM Cortex-M0 lithium battery, the value is 0.80. In our research, the
(2)ZigBee module lithium battery would not be charged until its voltage
(3)Sensors (dissolved oxygen, PH, temperature etc.) dropped under a specific level, and it should provide
the system with the rest energy just in case there is no
The power consumption of all the components is cal- sunshine for charging, namely, even when the battery
culated as shown in Table 2. Vin represents the input voltage is reduced to the charging boundary, it still per-
voltage, which is the ISEH system output voltage. The sists that appropriate energy as a backup should be
notation I stands for the average current. Especially, the considered, so the value of cc is set to 0.6.
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 6 of 12

Table 3 Comparison of rechargeable batteries

Sequence number Voltage/V Volumetric energy The number Self-discharge Memory effect Environmental
density/Wh/L of cycles rate/% M-1 protection
1 2.0 60 ~ 75 250 ~ 300 5 ~ 15 No Poisonous
2 1.2 110 ~ 130 300 ~ 700 15 ~ 30 Yes Poisonous
3 1.2 140 ~ 300 400 ~ 1000 25 ~ 35 Little Harmfulness
4 3.7 250 ~ 360 500 ~ 1000 5 ~ 10 No Harmfulness
5 3.7 300 ~ 460 500 ~ 1000 2~5 No Non-poisonous

Combined with the data from above, the battery cap- parallel to the horizontal axis is the charging threshold,
acity is calculated to be about 3190−4061 mAh (for the and the dashed black square waves represent the inten-
nominal voltage of a lithium-polymer battery). Since the sity of the sunshine. The blue curve is a reference, one
calculation is made with safety allowance, the battery that represents the continuous discharge of the lithium
capacity can be set to 4000 mAh. battery voltage without charging. As shown in the figure,
the lithium battery of the ISEH system discharges only
5 Simulation and implementation when the sunshine is insufficient, and it is charged once
In this section, the performance of the proposed ISEH during the whole 40 h. Relatively, the charge-discharge
system is simulated and compared with two types of cycle of the CWSS system is six times, which is far more
existing solar energy-harvesting systems [16, 30–32, 35], than that of the ISEH system. At the same time, it is
then we gave the design details of each module in the SOC also in a high level most of the period. As a matter
ISEH system. Finally, we have implemented an experi- of fact, the real weather condition cannot be regular like
mental circuit, as shown in Fig. 9, which was tested with that, so the number of the cycle is usually more than
output characteristics of the system. that shown in this simulation, which has a great influ-
ence on the life of lithium batteries.
5.1 Performance simulation Figure 4 shows the comparison result between the ISEH
Based on the analysis mentioned above, simulations are system and the other system which uses the lithium bat-
taken for demonstrating the differences between the tery as the main power supply (BMPS) and charges it only
ISEH system and the other systems. The parameters when its voltage is beneath a specific level. The parame-
used in the simulations are shown in Table 4. ters used in the simulations are also shown in Table 4. Ex-
It is worth noticing that the duty cycle is set to 1, cept for the red curve with “o” which represents the
which means that the node works all the time with full battery voltage of the BMPS system, the rest of the curves
power consumption, and the sunshine performs a regu- are the same as that shown in Fig. 3.
lar pattern like a square wave, which is unreal for the ac- As shown in Fig. 4, the BMPS system makes a little
tual weather condition. All the settings are adapted only improvement, because the charge-discharge cycle is less
for the reason of facilitating the contrast effect. Results
given in Fig. 3 are a comparison between the ISEH sys-
tem and the previous system which could charge when
the sunlight is sufficient (CWSS), which charges the bat-
tery as long as there is sufficient illumination.
In Fig. 3, the green line represents the lithium battery-
voltage curve of the ISEH system and the red line repre-
sents that of the CWSS system; the dashed red line

Table 4 Simulation parameters

Lithium battery 4000 mAh, 3.7 V
Power consumption of node 460 mW
Duty cycle 1
Total working hours 40 h, 6 h for a loop,
which includes 4 h
of efficient sunshine
and two overcast hours Fig. 3 Simulation result between the ISEH and CWSS system
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 7 of 12

Fig. 5 Charging profile of CN3063

Fig. 4 Simulation result between the ISEH and BMPS system

The lithium battery-charging process is generally di-
vided into the following three phases:
than that of the CWSS system shown in Fig. 3. But
whatever the weather is, the lithium battery is the main (1)Pre-charge phase
power source, which causes a waste of the solar energy (2)Constant current phase
and makes the discharge rate larger than that of the (3)Constant voltage phase
ISEH system, thus the number of charge-discharge cy-
cles is also more than that of the ISEH system. The first phase is carried out only when the lithium
battery voltage is very low. Usually, the second and the
5.2 Detailed design third phase occupy most of the charging process. The
According to the principle of the system composition, constant current of charging can be regulated externally,
each function module of the system is taken into de- and the continuous programmable charge current can
tailed circuit design. The main functional module reach 600 mA.
realization is described as follows. In the second step, the calculation formula of charging
current is given by

5.2.1 MPPT module ICH ¼ 1800 V=RISET ð2Þ

The MPPT module is based on chip MP2307 which uses
an autonomous tracking strategy to provide an upmost
power supply for the system. The input of the module is ICH - charging current, with unit A
the output voltage of the solar panel, and the output of RISET - the resistance from the ISET pin of the CN3063
it is the normal voltage. to ground, with unit Ω

For example, if the system requires a charging current

5.2.2 Charging module
of 500 mA, RISET can be calculated as follows:
This module is based on the charging management chip
CN3063 [38] which is produced by the Consonance RISET ¼ 1800 V=0:5A ¼ 3:6kΩ ð3Þ
Electronic Co. Ltd. The function of this module provides
an intelligent method to charge the lithium battery. To ensure good stability and temperature characteris-
The input of this module is from the MPPT circuit, and tics, RISET recommends the use of a metal film resistor
the output is connected to the lithium battery. CN3063 is with an accuracy of 1 %.

a single lithium battery-charging management chip, which When charging continues, the level of the CHRG pin
can be easily powered by a solar panel with a wide range drops down by an internal switch, which indicates that the
of input voltage. An on-chip 8-bit ADC can adjust the charging is in progress; otherwise, the pin is in a high-
charging current automatically. Its regulation voltage impedance state. This pin can be used to indicate whether
can be adjusted by an external resistor, and the charging the battery is being charged or not by an external LED.

current can also be programmed externally with a single When the charging is terminated, the level of the DONE
resistor. The charging process of CN3063 is drawn in pin drops down by an internal switch; otherwise, the pin is
Fig. 5 [38]. in a high-impedance state. The pin’s low level can be used
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 8 of 12

as the input signal to the RS trigger, and it turns the trigger a three-terminal regulator (OUTPUT) and the positive-
over and maintains the status, which ensures that the lith- going input which is provided by the MPPT output volt-
ium battery cannot be charged until its voltage drops age. When the positive-going input voltage is higher than
below the predefined level again. the reference, LM339 provides a high-level output, which
makes Q1 (9013) switch on and drives the pin G of the
5.2.3 Boost module Mosfet (U3) down; thus, U3 breaks over, and the system
There are some differences between the battery nominal selects the solar panel for power support. At the same time,
voltage and the system output voltage; therefore, it is ne- pin 1 of U4A is high, which makes ST2301 (U13) and
cessary to use a DC/DC boost circuit to make the conver- Mosfet (U6) both go off and then the output of the lithium
sion. GS1661 is a current mode boost DC/DC converter, battery is closed.
and it is available in SOT23-6 L package and provides On the other hand, when the positive-going input volt-
space-saving PCB for applications. The chip can be imple- age is beneath the reference, the output voltage of U4A
mented with fewer external components. should be low, and ST2301 (U13) switches on making
The proposed module can raise voltage from 3.3–4.2 V U6 break over. As a matter of fact, the lithium battery
to the system output voltage, and the output voltage can provides the power instead of the solar panel.
be determined by two external resistors
5.2.5 Charge control and over discharge protection
VOUT ¼ 0:6V  ðRe1 =Re2 þ 1Þ ð4Þ To prolong the life of the lithium battery, lengthening the
charge-discharge cycle as long as possible is undoubtedly
The peak current can also be determined by an exter-
a good approach. Li et al. used a microcontroller to realize
nal resistor Re3, the expression is
the charging progress, which is convenient and makes it
IOCP ¼ 48; 000 V=Re3 ð5Þ easy to change some strategies, and it merely seems
slightly more complicated for programming [35]. In this
paper, we deploy a sophisticated method using RS triggers
5.2.4 Control module of branch choosing to implement this function with less cost. Consequently, it
Since the ISEH system has two branches for power supply, is more stable compared with the crash possibility of the
a switching control circuit as shown in Fig. 6 is needed. microcontroller.
LM339 is a voltage comparator (U4A), and it has two in- The charge control and over discharge protection cir-
puts, the inverted input as a reference which is provided by cuit is shown in Fig. 7.

Fig. 6 Power supply circuit switching module

 Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 9 of 12

Fig. 7 Charge control and over discharge protection

The lifetime of the lithium battery determines the life to Table 5, the output of the RS trigger maintains the
of the entire system. Based on the premise of not affect- former state, and the system still continues to charge
ing the system work, reducing the charge-discharge cy- the battery.
cles of the battery is very helpful. Experimental results 
When the battery has been charged, the pin DONE of
show that the lithium battery can work for a long time 
CN3063 converts to low, and the pin 2S is also being
when its voltage is between 3.7–3.9 V, but if the voltage 
low. Simultaneously, the pin 2R remains high, so the RS
is lower than the threshold, it drops rapidly with dis- trigger outputs high levels, and the charging circuit is
charge going on. So the lithium battery should be 
disconnected. The pin DONE turns to being in a high-
charged at once when its voltage is under the threshold. 
Critically, if the battery voltage is lower than the danger impedance state once again, which makes the level of 2S

boundary, it may be eternally damaged, so discharge high, and 2R also maintains a high state, thus the RS
should be stopped immediately, and the system has to trigger remains in a Q0 state of high level, which means
do something to avoid this from happening [35]. the circuit is out of the charging process.
According to the circuit in Fig. 7, when the battery RS trigger realizes an intelligent charging strategy.
voltage is lower than the threshold (whose value can be When the lithium battery voltage falls below the thresh-
adjusted by the resistor R14 and R18), the output of old, the charging circuit will switch on. Subsequently,
 when the battery is fully charged, the circuit will discon-
U4B in LM339 and the pin 2R of RS trigger (U10) are
both low. Since the battery is not fully charged, the pin nect at once. It would not be turned on until the battery

DONE of CN3063 is in a high-impedance state accord-

ing to its data sheet. Therefore, the pin 2S of U10 is
high. Based on the function table of RS trigger as shown Table 5 Function table of RS Trigger
in Table 5, the output Q of RS trigger should be low, Inputs Output
thus the Mosfet U9 is conducted, and the system starts – –
S R Q
to charge the battery. L L H
Once the battery voltage is higher than the threshold
 L H H
with continuous charging, the level of 2R changes from
 H L L
low to high, and 2S remains at a high level at the same
H H Q0
time because charging has not finished yet. According
 Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 10 of 12

Fig. 8 The RS trigger sequence diagram

voltage drops beneath the threshold again. All the de- As shown in Table 2, the power consumption of one sen-
scriptions in time sequence are shown in Fig. 8. sor node is 460 mW. We can get two different operating
In addition, the system also has an over-discharged points in the conditions of solar power or lithium battery-
protection module. When the lithium battery voltage is branch energy supply, as shown in Fig. 10a. The reason
lower than the protection (the value can be adjusted by causing the operating points to be different is that the out-
two resistors), U12 (ST2301) is disconnected, which cuts put voltage of the two branches have a slight difference in
off the lithium battery power output to protect the bat- the circuit. Whether the power is provided by the solar or
tery from overdischarge. the lithium branch, the power efficiency (as shown in
Fig. 10b) is higher than 60 %. It is worth mentioning that
5.3 Experimental results the power efficiency is up to 80 % with the lithium battery
F9
A circuit board for testing is made as shown in Fig. 9. The as a power supply. From Fig. 10b we can find that the effi-
ISEH system is constructed based on the board, which ciency of the solar power branch is lower than that of the
could provide uninterrupted power supply for the WSN lithium branch, and at the same time, the former is more
nodes. The system is tested under daylight illumination, sensitive with the load resistance. The reasons for this
and the experimental results show that the system can phenomenon can mainly be concluded in two parts. The
switch the power supply branch automatically and work first one is that the voltage difference between input and
stably. When the lithium battery voltage drops underneath output of the solar branch is larger than that of the battery
the predefined level, the lithium battery can be charged as branch, and the second is that the former components
soon as possible if the illumination is sufficient. cause more power loss than that of the latter.

Fig. 9 Circuit board

Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 11 of 12

Competing interests
The authors declare that they have no competing interests.

Acknowledgements
This work was supported by the Hunan Provincial Natural Science
Foundation of China under Grant 14JJ5009, the Post Doctoral Foundation of
Central South University, China, and the National Natural Science Foundation
of China (61272495, 61379153, and 61401519).

Received: 7 January 2015 Accepted: 11 June 2015

References
1. ZhouY, Yang X, Guo X, et al. A design of greenhouse monitoring & control
system based on ZigBee wireless sensor network. IEEE. Int. Conf. WiCOM,
(IEEE, Shanghai, 2007), pp. 2563–2567
2. Ahonen T, Virrankoski R, Elmusrati M. Greenhouse monitoring with wireless
sensor network. IEEE/ASME. Int. Conf. on MESA, (IEEE/ASME, Beijing, 2008),
(a) pp. 403–408
3. Nasser N, Zaman A, Karim L, et al. CPWS: An efficient routing protocol for
RGB sensor-based fish pond monitoring system. IEEE. 8th. Int. Conf. WiMob,
(IEEE, Barcelona, 2012), pp. 7–11
4. O’Flyrm B, Martinez R, Cleary J, et al. Smart Coast: a wireless sensor network
for water quality monitoring. IEEE. 32nd. Conf. LCN. 2007;815–816.
5. GS Arumugam, T Ponnuchamy, EE-LEACH: development of energy-efficient
LEACH Protocol for data gathering in WSN. EURASIP. J. Wireless. Commun.
Netw. 1, 76 (2015). doi:10.1186/s13638-015-0306-5
6. I Butun, I Ra, R Sankar, PCAC: power-and connectivity-aware clustering for
wireless sensor Networks. EURASIP. J. Wireless. Commun. Net. 1, 83 (2015).
doi:10.1186/s13638-015-0321-6
7. G Anastasi, M Conti, M Di Francesco et al., Energy conservation in wireless
sensor networks: a survey. Ad. Hoc. Netw. 7(3), 537–568 (2009)
8. RJM Vullers, RV Schaijk, HJ Visser et al., Energy harvesting for autonomous
wireless sensor networks. IEEE. Solid State Circuits. Mag. 2(2), 29–38 (2010)
9. F Akhtar, MH Rehmani, Energy replenishment using renewable and
traditional energy resources for sustainable wireless sensor networks: a
review. Renew. Sustain. Energy Rev. 45, 769–784 (2015)
10. VA Boicea, Energy storage technologies: the past and the present. Proc.
(b) IEEE. 102(11), 1777–1794 (2014)
11. Z Xu et al., Electrochemical supercapacitor electrodes from sponge-like
Fig. 10 Output characteristics of the ISEH system. a Power graphenenano architectures with ultrahigh power density. J. Phys. Chem.
conversion efficiency versus load resistance. b Output power versus Lett. 3(20), 2928–2933 (2012)
load resistance 12. W Waag, S Käbitz, DU Sauer, Experimental investigation of the lithium-ion
battery impedance characteristic at various conditions and aging states and
its influence on the application. Appl. Energy. 102, 885–897 (2013)
13. M Broussely, P Biensan, F Bonhomme et al., Main aging mechanisms in Li
6 Conclusions ion batteries. J. Power. Sources. 146(1), 90–96 (2005)
14. J Vetter, P Novák, MR Wagner et al., Ageing mechanisms in lithium-ion
In this paper, a novel intelligent solar energy-harvesting batteries. J. Power. Sources. 147(1), 269–281 (2005)
system is designed by using an MPPT circuit. Hardware, 15. M Ecker, JB Gerschler, J Vogel et al., Development of a lifetime prediction
instead of software, is used for charging management of model for lithium-ion batteries based on extended accelerated aging test
data. J. Power. Sources. 215, 248–257 (2012)
the lithium battery, which can enhance the robustness of 16. Jiang X, Polastre J, Culler D. Perpetual environmentally powered sensor
the system greatly. Analyses based on power supply networks. IEEE. 4th. Int Symp. IPSN, (IEEE, Los Angeles, 2005), pp. 463–468
requirements are made for WSN nodes in IOT. The 17. Xiang Y, Pasricha S. Run-time management for multicore embedded systems
with energy harvesting. 2015. doi:10.1109/TVLSI.2014.2381658
system can afford a stable power supply with 5-V output 18. Dutta P, Hui J, Jeong J, et al. Trio: enabling sustainable and scalable outdoor
voltage through a standard USB interface. Lithium wireless sensor network deployments. Proc. 5th. Int. Conf. Info. Process.
battery-charging strategy can also ingeniously avoid the Sensor. Netw, (ACM, Nashville, TN, 2006), pp. 407–415
19. A Berger, LB Hörmann, C Leitner et al., Sustainable energy harvesting for
charge-discharge cycle a lot, and thus the lifetime of the robust wireless sensor networks in industrial applications (Sensor Applications
lithium battery can be greatly extended. Experimental Symposium SAS, Zadar, 2015)
results demonstrate that the system can switch the 20. Dall’Ora R, Raza U, Brunelli D, et al. SensEH: From simulation to deployment
of energy harvesting wireless sensor networks. IEEE. 39th. Conf. Local.
power supply branch automatically. When the voltage of Comput. Netw. Workshops, (IEEE, Edmonton, AB, 2014), pp. 566–573
lithium battery drops below the predefined level, it can 21. V Sharma, U Mukherji, V Joseph et al., Optimal energy management
be charged properly. The system performs stably and policies for energy harvesting sensor nodes. IEEE. Trans. Wireless. Commun.
9(4), 1326–1336 (2010)
safely with high reliability, high efficiency, low-power 22. Ramasur D, Hancke G P. A wind energy harvester for low power wireless
loss, and simple composition. sensor networks. IEEE. Int. Conf. I2MTC, (IEEE, Graz, 2012), pp. 2623–2627
Li and Shi EURASIP Journal on Wireless Communications and Networking (2015) 2015:179 Page 12 of 12

23. JA Khan, HK Qureshi, A Iqbal, Energy management in wireless sensor

networks: a survey. Comput. Electr. Eng. 41, 159–176 (2015)
24. Y Gao, G Sun, W Li et al., Wireless sensor node design based on solar
energy supply. IEEE. 2nd. Int. Conf. PEITS 3, 203–207 (2009)
25. Naveen K V, Manjunath S S. A reliable ultra-capacitor based solar energy
harvesting system for wireless sensor network enabled intelligent buildings.
IEEE. 2nd. Int. Conf. IAMA, (IEEE, Chennai, 2011), pp. 20–25
26. Chuang WY, Lee CH, Lin CT, et al. Self-sustain wireless sensor module. IEEE
Int Conf Internet Things Green Computing Commun Cyber Physical Soc
Computing 2014;288–291
27. Taneja J, Jeong J, Culler D. Design, modeling, and capacity planning for
micro-solar power sensor networks. Proc. 7th. IEEE. Int. Conf. IPSN, (IEEE
Computer Society, St. Louis, Missouri, 2008), pp. 407–418
28. Alberola J, Pelegri J, Lajara R, et al. Solar inexhaustible power source for
wireless sensor node. IEEE. Proc. IMTC, (IEEE, Victoria, BC, 2008), pp. 657–662
29. Bhuvaneswari P T V, Balakumar R, Vaidehi V, et al. Solar energy harvesting
for wireless sensor networks. IEEE. 1st. Int. Conf. CICSYN, (IEEE, Indore, 2009),
pp. 57–61
30. Isaacson M J, Holl. Sworth R P, Giampaoli P J, et al. Advanced lithium ion
battery charger. IEEE. 15th. BCAA, (IEEE, Long Beach, CA, 2000), pp. 193–198
31. F Ongaro, S Saggini, P Mattavelli, Li-ion battery-supercapacitor hybrid storage
system for a long lifetime photovoltaic-based wireless sensor network. IEEE.
Trans. Power. Electronics. 27(9), 3944–3952 (2012)
32. H He, R Xiong, X Zhang et al., State-of-charge estimation of the lithium-ion
battery using an adaptive extended Kalman filter based on an improved
Thevenin model. IEEE. Trans. Veh. Technol. 60(4), 1461–1469 (2011)
33. K Takahashi, M Saitoh, N Asakura et al., Electrochemical properties of lithium
manganese oxides with different surface areas for lithium ion batteries.
J. Power. Sources. 136(1), 115–121 (2004)
34. YJ Liu, XH Li, HJ Guo et al., Electrochemical performance and capacity
fading reason of LiMn2O4/graphite batteries stored at room temperature.
J. Power. Sources. 189(2), 721–725 (2009)
35. X Li, Y Wu, G Li et al., Development of wireless soil moisture sensor base on
solar energy. Trans. CSAE. 26(11), 13–18 (2010)
36. D Brunelli, C Moser, L Thiele et al., Design of a solar-harvesting circuit for
batteryless embedded systems. IEEE. Trans. Circuits. Syst I: Regul. Pap.
56(11), 2519–2528 (2009)
37. T Esram, PL Chapman, Comparison of photovoltaic array maximum power
point tracking techniques. IEEE. Trans. Energy. Conversion. EC. 22(2),
439 (2007)
38. CN3063 datasheet, Shanghai Consonance Electronics Co. Ltd.,
http://www.consonance-elec.com/pdf/datasheet/DSE-CN3063.pdf.

Submit your manuscript to a

journal and benefit from:
7 Convenient online submission
7 Rigorous peer review
7 Immediate publication on acceptance
7 Open access: articles freely available online
7 High visibility within the field
7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com