Received: 17 April 2020 Revised: 6 June 2020 Accepted: 24 June 2020

DOI: 10.1002/er.5726

RESEARCH ARTICLE

Small-scale experimental study on the optimisation of a

rooftop rainwater energy harvester using electromagnetic
generators in light rains

Bin Bao1 | Quan Wang1,2

1
Department of Mechanics and Aerospace Engineering, Southern University of Science and Technology, Shenzhen, China
2
Department of Civil and Environmental Engineering, Shantou University, Shantou, China

Correspondence
Quan Wang, Department of Mechanics Summary
and Aerospace Engineering, Southern This research proposes a new rooftop rainwater energy harvester using electro-
University of Science and Technology,
magnetic generators, which demonstrates better rain-energy harvesting perfor-
Shenzhen 518055, China.
Email: [email protected] mance than the traditional rainwater energy harvesters that use
electromagnetic generators under light rain conditions. Conventionally, the
Funding information
rainwater collected from the roof of a building or a residential attic using the
Shenzhen Science and Technology
Planning Project, Grant/Award Number: installed roof gutter system is drained directly along a pipe to drive water tur-
JCYJ20170817111535167; Key Scientific bines combined with electromagnetic generators for generating electricity.
Research Projects of Universities in
Such an energy harvesting method has high energy harvesting efficiency under
Guangdong Province, Grant/Award
Number: 2018KZDXM064 heavy rain conditions but is inefficient under light rain conditions. To improve
the energy harvesting efficiency in light rains, a pipe-connected tank including
a passive rainwater buffer is introduced and fixed below the rooftop gutters in
the proposed structure. The passive rainwater buffer is attached above the
pipe-connected tank. After collecting a certain amount of rainwater from the
gutters, the passive rainwater buffer automatically releases the rainwater peri-
odically into the pipe-connected tanks. The released rainwater results in a large
flow along the pipe, which can drive water turbines with electromagnetic gen-
erators to work at a high efficiency for a short period of time. A small-scale
experimental prototype was built for validating the energy harvesting perfor-
mance of the proposed structure. The results show that the intermittent high-
efficiency power generation of the proposed structure is better than the contin-
uous low-efficiency power generation of the traditional structure during light
rains. The proposed structure can charge a 680 μF capacitor with at least nine
times the electrical energy that is stored in a conventional capacitor for a rain-
water flow rate of less than 300 mL/min in the same amount of time. However,
with the increase in the rainwater flow, the energy harvesting performance gap
between the proposed structure and the traditional one gradually narrows.

Int J Energy Res. 2020;1–19. wileyonlinelibrary.com/journal/er © 2020 John Wiley & Sons Ltd 1
2 BAO AND WANG

Highlights
1. A new rooftop rainwater electromagnetic energy harvester is proposed;
2. A pipe-connected tank including a passive rainwater buffer is fixed below
the rooftop gutters;
3. The proposed structure can charge a capacitor with at least nine times as
much electrical energy as the conventional one in light rain conditions.

KEYWORDS
electromagnetic generator, energy harvesting, rain energy, rainwater buffer, rooftop, turbine

1 | INTRODUCTION and poly (3,4-ethylenedioxythiophene)/PtCo, to yield elec-

tricity in a rainy atmosphere. An optimised CP/PtCo hybrid
The availability of water resources in several regions such generates a maximum current of 2.06 μA/droplet and volt-
as India1 and Jordan2 is significantly affected by climate age of 97.43 μV/droplet. Bao et al20 proposed a piezoelectric
change, the rapid development of urban areas, and the pop- hydro-energy harvester using a special cylindrical container
ulation growth in recent decades. A rainwater harvesting for low water-head and flow-rate applications.
system can provide an alternative sustainable water Another process of energy harvesting from rainfall
resource.3 It lessens the strain on the current water supply relies on rotational generators such as water turbines.21,22
system and helps to protect the quality of surface water. It In the conventional rooftop rainfall energy harvesters,
is suitable for all types of buildings ranging from residential water turbines are extensively used as rotational genera-
to commercial and industrial structures. Typically, the roof- tors.23 Normally, water turbines24 are divided into
top water4 collected by rainwater harvesting systems is used impulse turbines (ie, Pelton, Turgo, and Crossflow tur-
for non-potable demands (such as industrial processing) bines) and reaction turbines (ie, Francis, Propeller, and
and potable demands (such as cooking).5 Amongst these Kaplan turbines). According to different water-head con-
demands, energy harvesting from rainfall during the rain- ditions (corresponding to the height of buildings) and
water harvesting process has attracted significant attention other real-time environmental conditions, different water
in recent years. Rainwater energy harvesting can not only turbines were developed and applied for rooftop rain
convert the mechanical energy from falling raindrops into energy harvesting. The potential kinetic energy of the
electricity for powering sensors and other electronics moving rainwater changes multiple times as the rainwa-
devices but can also effectively solve ground collapse prob- ter falls from the sky and the gutters of the installed roof
lems caused by the impact of rainwater. gutter systems on the building roof or residential attic
Several methods were developed to collect the poten- collects and transports the rainwater, which reaches and
tial energy of rain in recent years,6,7 which can be divided rotates the water turbine. The water turbine is combined
into two categories. Piezoelectric materials are widely used with an electromagnetic generator to produce electric
for vibration energy harvesting.8-13 One type of harvester energy. Electromagnetic systems are effective and com-
relies on the rainwater-induced kinetic-to-electric energy monly used for energy harvesting.25,26 Based on the theory
transfer through piezoelectric materials attached to elastic of electromagnetism, the rotating turbine is connected to a
structures. The most common method utilises the ambient stator and rotor for producing electricity. The generator
vibration energy induced by direct impact of raindrops, converts the mechanical energy from the turbine into elec-
which is converted into electrical energy through the pie- trical energy. Then, the electrical energy is stored in a bat-
zoelectricity of piezoelectric materials.14-16 Several tery after the AC output is converted to DC output using
researchers have already studied this type of the rain energy harvesting circuits. Several researchers have investi-
energy harvester.17 Jellard et al18 investigated the key gated and developed new water turbines or other methods
parameters which affect energy transfer efficiency between for scavenging the rooftop rain energy in different practical
the incident water droplets and piezoelectric cantilever applications.27-31 Martin et al32 studied the energy
structures made of stainless steel foil coated with piezo- harvesting performed from rainwater using a Pico turbine.
electric polyvinylidene fluoride trifluoroethylene. Wang The generation scheme of the Pico turbine was designed
et al19 experimentally developed a category of film-type and developed using MATLAB/SIMULINK software.
rain energy harvesters from conductive polymer (CP)/PtCo Bhargav et al33 utilised the potential energy of a high rain-
hybrids, including polypyrrole/PtCo, polyaniline/PtCo, water head in a tall building to produce electrical energy
BAO AND WANG 3

using a micro hydro turbine. The power produced was harvester, a small pipe-connected tank with a passive rain-
1.54 W using a single turbine. The exergy wasted was water buffer is fixed between the rooftop gutters and the
36.75 W, indicating that a considerable amount of energy rainwater pipe, as shown in Figure 1B. The rainwater buffer
can still be extracted. In addition, piezoelectric plucking is attached above the pipe-connected tank. When the rain-
transducers34 can also be employed at the core of these water buffer container is full, it automatically releases all
rotational generators for rainwater energy harvesting. the rainwater to the pipe-connected tank. Assuming the
However, the mechanical energy from the water turbine rainwater flow rate is constant, the rainwater buffer periodi-
and the corresponding AC output of electromagnetic gener- cally releases the rainwater to the pipe-connected tank. The
ators vary according to the heaviness of rain. The rainwater rainwater from the pipe-connected tank flows along the
energy harvesting process using water turbines exhibits a attached rainwater pipe and reaches the turbine, resulting
high energy harvesting efficiency under heavy rain condi- in its rotation. As the rainwater buffer periodically supplies
tions; however, the energy harvesting efficiency is low sufficient rainwater under the light rain conditions, the
under light rain conditions. Only a few studies address the water turbine achieves a high energy-collection efficiency
low-efficiency problem due to light rain conditions. To for a period of time. When the turbine is rotated, the elec-
improve the power generation efficiency of the rainwater tromagnetic generator converts the mechanical energy from
energy harvesting process using water turbines combined the turbine into AC output electrical energy. The AC output
with electromagnetic generators under light rain condi- electrical energy is converted to DC output electrical energy
tions, this research proposes a new rooftop rainwater by the energy harvesting circuits for charging a battery or
energy harvester using water turbines with electromagnetic powering electrical loads.
generators. It is ideal for optimising all the traditional roof- The rainwater buffer is composed of a square tank that
top rainfall energy harvesting methods which use water tur- includes a mass block. A rotating shaft passes through the
bines with electromagnetic generators. In the proposed centre of the tank. The mass block inside the tank is used
structure, a small pipe-connected tank with a passive rain- for adjusting the centre of gravity of the tank. Owing to the
water buffer is fixed between the rooftop gutters and the mass block, the rainwater buffer is tilted when it is empty.
rainwater pipe. The passive rainwater buffer having a self- Figure 2 shows the diagram of the rainwater buffer. The
flip function35 is attached above the pipe-connected tanks. detailed parameters of the rainwater buffer are listed in
The rain energy harvesting performance of the proposed Table 1 and are used in the following experiments. The
and traditional structures was experimentally compared. tank flips around the shaft on bearings to release the rain-
The results demonstrated that the intermittent high- water into the pipe-connected tank when the rainwater
efficiency power generation of the proposed structure is bet- increases to reach the full capacity of the tank. Then, the
ter than the continuous low-efficiency power generation of tank returns to its original inclined state.
the traditional structure in light rain conditions. This article
has five sections. After a brief introduction on the signifi-
cance of the work, Section 2 describes the proposed rooftop 3 | MODELLING OF THE
rainfall energy harvester and compares if with the tradi- PROPOS ED STRUCTUR E
tional methods. Section 3 discusses the working principles
of the proposed structure and the mechanism of the passive 3.1 | Working principles of the proposed
rainwater buffer. Section 4 presents the experimental vali- structure
dation of the proposed structure. Finally, the conclusions
are presented in Section 5. Figure 3 shows the working principle of the proposed
rainfall energy harvester. A turbine cycle can be divided
into waiting and working times. The proportion of each
2 | OVERVIEW OF THE PROPOSED cycle at different times depends on the heaviness of rain.
ENERGY HARVESTER During the waiting time, the water turbine is inactive
and waits for the rainwater flow from the pipe-connected
Figure 1 shows the diagrams of the proposed and tradi- tank. Conversely, during the working time, the water tur-
tional rooftop rainfall energy harvesters using water tur- bine rotates and the connected electromagnetic generator
bines with electromagnetic generators. Traditionally, the generates electricity. Specifically, the waiting time
rainwater collected from the building roof or residential includes three waiting phases. The first waiting phase
attic using the installed roof gutter systems is drained denotes the state at which the turbine is stationary when
directly to a pipe for driving the water turbines attached to no rainwater flows into the rainwater buffer. The second
the electromagnetic generators for generating electricity, as waiting phase denotes the state at which the rainwater
shown in Figure 1A. In the proposed rooftop rainfall energy gradually flows into the rainwater buffer before it is full.
4 BAO AND WANG

F I G U R E 1 A, Traditional
rooftop rain energy harvester
with water turbines which drive
the electromagnetic
generators; B, proposed rooftop
rain energy harvester with a
passive rainwater buffer fixed
above a small pipe-
connected tank [Colour figure
can be viewed at
wileyonlinelibrary.com]

dΦ
The third waiting phase denotes the state at which the rain- E = −N , ð1Þ
dt
water from the buffer is poured into the pipe-connected
tank and the water turbine gets ready to work owing to suf- where the negative sign describes the direction in which
ficient rainwater flow. The working time of the turbine cycle the induced emf drives the current around a circuit.
includes two phases. The first working phase denotes the However, this direction is easily determined using Lenz's
state when the running rainwater flows through the pipe- law. N denotes the number of turns of a coil. dΦ /dt
connected tank, and through the pipe, the water reaches denotes the rate of change of the number of magnetic
and spins the water turbine. The second working phase flux lines passing through the surface.
denotes the state when the water turbine continues to rotate The magnitude of the current induced in the coil
owing to rotational inertia, even after the rainwater in the I can be presented as:
pipe-connected tank has run out. The different phases trans-
form in to one turbine cycle as shown in Figure 3. jE j
I= , ð2Þ
The operating principle of the electromagnetic gener- R
ator incorporates the concept of electromagnetic induc-
tion. According to Faraday's law, the induced where R denotes the resistance of the coil.
electromotive force (emf) E or the voltage generated by a Therefore, the mechanical energy from the turbine
conductor or coil moving in a magnetic field can be can be converted to AC output electrical energy by the
determined by: electromagnetic generator according to Faraday's law. In
BAO AND WANG 5

FIGURE 2 Diagram of the passive rainwater buffer [Colour figure can be viewed at wileyonlinelibrary.com]

TABLE 1 Parameters of the passive rainwater buffer This study focused on enhancing the energy harvesting
Parameters Value performance in light rain conditions.
ts 0.004 m
ti 0.003 m
3.2 | Function analysis of the passive
a 0.03 m
rainwater buffer
hs 0.14 m
am 0.02 m The global centre of gravity of the rainwater buffer con-
hg 0.06 m tainer gradually goes up with the increase in the rainwa-
ρc 955 kg/m3 ter level. When the overall centre of gravity rises above
the axis of rotation and is not in a symmetrical plane, the
ρm 1358 kg/m3
rainwater in the container will be in an unstable state;
ρwater 1000 kg/m3
thus, the container easily overturns and empties all the
rainwater. Figure 4 shows the half-cut cross section of
this study, we only installed commonly used water tur- the container when it begins to overturn owing to the
bines and electromagnetic generators in the proposed unstable state. Assuming the inclination angle of the con-
structure as the focus of this research was not on the opti- tainer is θ, the angle between the water surface and the
misation of water turbines or electromagnetic generators. line of the container opening will be equal to θ.
6 BAO AND WANG

F I G U R E 3 Working
principle of the proposed
structure [Colour figure can be
viewed at
wileyonlinelibrary.com]

When 0 ≤ θ < tan−1([hs-am]/2a), as shown in container on the shaft O, respectively. Here, g = 9.81 m/
Figure 4A, the rainwater block A can be divided into s2. Through calculations, MrwA1, MrwA2, and MrwA3 can be
three block parts: A1, A2, and A3. Assuming that the tor- correspondingly obtained as follows:
que required to pour water out by tilting the container is
" r
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
positive, the torque Mrw exerted by the rainwater in the am2  am2
buffer container on the shaft O in this condition can be M rw
A1 = g 2ρwater aam hg −
2
+ a− sin
2 2
calculated as:



 #
−1 hs− am a− am am 
A1 + M A2 + M A3 , 0 ≤ θ < tan
M rw = M rw ð3Þ
rw rw
, θ −tan −1 2
−4a amρwater sinθ hg −
2
:
2a hg − am 2
2

where MrwB1, MrwB2, and MrwB3 denote the torque exerted ð4Þ
by the rainwater blocks A1, A2, and A3 in the buffer
BAO AND WANG 7

22 0 0 1 1 3 3
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
66 B −1B a C C 2a 7 7
6 4 ðhs −hgÞ + a2 cos@cos @qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiA − θA −
2
5 7
6 ðhs− hgÞ + a 2 cosθ 7
6 2
7
6 2 0 0 1 1 3 7
6 7
6 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi 7
6 6 7
6 + 4 ðhs −hgÞ2 + a2 cosB @ cos − 1 BqffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiC
@
a
A −θ
C7
A 5 7
6 7
6 ð hs −hg Þ 2
+ a 2 7
6 7
6 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

  7
4 π hs −hg −2atanθ 5
+ ðhs −hg −2atanθÞ2 + a2 sin + tan − 1 −θ
2 a
A3 = 4ρwater a gtanθ
M rw : ð6Þ
3
3

torque Mrw exerted by the rainwater in the buffer con-

A2 = 4a gρwater sinθðhs −2atanθ −amÞ
M rw 2

 tainer on the shaft O in this condition can be pres-
hs am ð5Þ
 + − atanθ −hg : ented as:
2 2



−1 hs −am hs
M rw = M rw
B , tan ≤ θ < tan − 1 , ð7Þ
Similarly, when tan−1([hs-am]/2a) ≤ θ < tan−1(hs/ 2a 2a
2a), the rainwater block is B, as shown in Figure 4B. The
where MrwB can be calculated as:

2
 3
1
6 −4a ρ 2
water sinθð hs− 2atanθ Þ hg − hs + atanθ 7
6 2 7
6 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

 7
6 a 7
6 2ðhs −hg −atanθÞsinθ − ðhg− hs + 2atanθÞ2 + a2 sin θ −tan −1 7
6 hg −hs + 2atanθ 7
6 + 4ρ 7
water a tanθ 7: ð8Þ
rw 3
M B = g 6
6 3 7
6 0 0 11 7
6 am 7
6 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi


ffi a− 7
6 am 2 am 2 B −1 B 2 CC 7
4 + 2ρwater aam  hg− 2 + a − 2 sin@θ −tan @
2
am A A 5
hg−
2

When tan−1(hs/2a) ≤ θ < π/2-tan−1(am/[hs-am]), the

rainwater block is C, as shown in Figure 4C. The torque
Mrw exerted by the rainwater in the buffer container on
the shaft O in this condition can be given as:

  am 
−1 hs π
M rw = M rw
C , tan ≤ θ < −tan − 1 , ð9Þ
2a 2 hs −am

where MrwC can be calculated as:

2  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

 3

2 2 cos θ −tan −1
hs −hg hs 2 sin θ −tan −1
a
6 2 ð hs −hg Þ + a − − hg 2
+ a 7
6ρ a sin ðθ Þ hg 7
6 water ahs ctanθ
2
7
6 3 7
M rw = g 6 0 0 1 1 7: ð10Þ
C 6 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
am 7
6

a − 7
6 am 2 am 2 B −1 B 2 CC 7
4 + 2aρwater am  hg− 2 + a − 2 sin@θ −tan @
2
am A A 5
hg−
2
8 BAO AND WANG

FIGURE 4 A, Half-cut cross section of the rainwater buffer container when 0 ≤ θ < tan-1([hs-am]/2a); B, half-cut cross section of the
rainwater buffer container when tan-1([hs-am]/2a) ≤ θ < tan-1(hs/2a); C, half-cut cross section of the rainwater buffer container when tan-
1
(hs/2a) ≤ θ < π/2-tan-1(am/[hs-am]); D, half-cut cross section of the rainwater buffer container when π/2-tan-1(am/[hs-am]) ≤ θ < π/2
[Colour figure can be viewed at wileyonlinelibrary.com]

π  am  π
−1
When π/2-tan−1(am/[hs-am]) ≤ θ < π/2, the rainwa- D , −tan
M rw = M rw ≤θ< , ð11Þ
2 hs− am 2
ter block is D, as shown in Figure 4D. The torque Mrw
exerted by the rainwater in the buffer container on the
shaft O in this condition can be calculated as: where MrwD can be calculated as:

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi    qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   
ρwater agðhs −amÞ 2
2 ðhs −hgÞ2 + a2 cos θ −tan − 1 hs −a hg − hssinθ
− am
− ðhg − amÞ2 + a2 sin θ −tan − 1 hg−aam
M rw
D =  :
tanθ 3
ð12Þ
BAO AND WANG 9

if it is full under this condition. With the increase in the

height hg, the shaft gradually approaches the opening of
the buffer. The work W performed by the rainwater on
the shaft becomes negative, which denotes that it is diffi-
cult to overturn the buffer. When hg is less than 0.06 m,
the work W performed by the rainwater on the shaft is
significantly greater than zero according to Equation (14).
The container will be overturned before it is full in this
situation. With the decrease in the height hg, the shaft
gradually approaches the bottom of the buffer. The work
W performed by the rainwater to the shaft remains posi-
tive and becomes larger, which denotes that the buffer is
easy to overturn. When the parameter hg is less than
0.05, we can observe that the torque is greater than zero
in the entire domain of the inclination angle
(0 < θ < π/2), which implies that the opening of the
buffer is in the downward direction. In this case, the
F I G U R E 5 Torque M with respect to the inclination angle θ
when the rainwater buffer container overturns under different
buffer is obviously unable to collect rainwater and there-
heights hg of the shaft above the bottom of the buffer container fore cannot be used for energy harvesting. For ensuring
[Colour figure can be viewed at wileyonlinelibrary.com] that the container is nearly full when overturned, the
height hg of the rotating shaft is set to 0.06 m, which can
result in the work performed by the rainwater on the
In summation, the torque Mrw exerted by the rainwa- shaft to be marginally greater than zero. In the conducted
ter in the buffer container on the shaft O when experiments, the height hg was set to be equal to 0.06 m.
0 ≤ θ < π/2 can be obtained as:
8

>
> − 1 hs−am 4 | EXPERIMENTAL
> A1
> M rw
+ M rw
+ M rw
,0 ≤ θ < tan
>
>
A2 A3
2a INVESTIGATION
>
>


>
>
> − 1 hs−am hs
< M rw
B , tan ≤ θ < tan − 1 4.1 | Experimental set up
rw
M = 2a 2a : ð13Þ
>
  am 
>
> hs π
>
>
rw
M C , tan − 1
≤ θ < −tan − 1
>
> 2a 2 hs −am Figure 6 shows a small-scale experimental set-up of the
>
>  am 
>
> π π proposed rooftop rainfall energy harvester using a water
: M rw , −tan − 1
≤θ<
D
2 hs −am 2 turbine with an electromagnetic generator. Water is
pumped up from a pool to a higher level for simulating
The rainwater in the buffer container is released the rainwater flow from the gutters at the rooftop. The
when the work W performed by the rainwater on the rainwater buffer was manufactured by 3D printing. The
rotating shaft is greater than zero, assuming that the fric- parameters of the buffer container are listed in Table 1.
tion torque of the shaft is equal to zero. W can be calcu- The parameters of the used generator are listed in
lated as: Table 2. An oscilloscope is used to measure the AC out-
put voltage from the electromagnetic generator or the DC
ðπ
2 output voltage from the energy harvesting circuit. For the
W= M rw dθ > 0: ð14Þ
0 experiments using the traditional structure, the water
flow was directed into the pipe-connected tank. The
energy harvesting performance was compared between
Figure 5 shows the torque Mrw with respect to the the traditional structure and the proposed structure
inclination angle θ when the rainwater buffer container through experiments, the details of which are presented
flips over at different heights hg of the shaft above the below.
bottom of the buffer container. The height hg of the shaft It should be noted that the flow rate of the water from
significantly affects the flipping state of the container. the pump in the experiments is used to represent the
When hg is greater than 0.06 m, the work W performed rainwater flow rate from the gutters to simplify the exper-
by the rainwater on the shaft is less than zero according iment. In practice, the rainwater flow rate from the gut-
to Equation (14). The container will not be inclined even ters is primarily determined by both the amount of
10 BAO AND WANG

F I G U R E 6 Experimental set up
[Colour figure can be viewed at
wileyonlinelibrary.com]

TABLE 2 Parameters of the electromagnetic generator and rainfall and the surface area of the rooftop. More rainwa-
energy harvesting circuit ter can be collected in the gutter system by enlarging the
surface area of the rooftop. Receiving a larger amount of
Parameters Notation Value or size
rainfall and/or increasing the surface area of the rooftop
Deg Diameter of the generator 24 mm
can improve the rainwater flow rate from the gutters.
Heg Height of the generator 12.5 mm
Therefore, the rainwater flow rate from the gutters does
leg Length of the output shaft of the 7 mm not denote the amount of the rainfall. For convenience,
generator
we considered that the increase in the water flow rate
Rated voltage range of the generator 1.5-9 V
from the pump in the experiments represents an increase
Veg Test voltage of the generator 3V
in the amount of rainfall for a certain rooftop area.
Ieg Test current of the generator 10 mA
reg Rotational speed of the generator 1500 rpm
Diode used in the energy harvesting 1 N4007 4.2 | Discussion on experimental results
circuit
C Capacitor used in the energy 680 μF, 10 V Figure 7 shows the AC output voltage curves of the elec-
harvesting circuit
tromagnetic generator in the traditional and proposed
BAO AND WANG 11

F I G U R E 7 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
100 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is
100 mL/min

structures when the rainwater flows at an ultra-low flow in the traditional structure achieved an average value of
rate of 100 mL/min. As shown in Figure 7A, the mea- approximately 0.5 V in the investigated time domain, as
sured AC output voltage of the electromagnetic generator shown in Figure 8A. In contrast, in the proposed struc-
in the traditional structure is approximately 0 V in the ture, four turbine cycles were performed in the investi-
investigated time domain; thus, it can be noted that the gated time domain, as shown in Figure 8B. One turbine
electromagnetic generator in the traditional structure cycle occupied approximately 97 seconds. The working
does not work at a flow rate of 100 mL/min. The reason time within one turbine cycle was approximately
for this phenomenon is that the rainwater flow rate of 27 seconds. Therefore, the water turbine worked effi-
100 mL/min does not have sufficient impact power to ciently for 114 seconds at a flow rate of 200 mL/min in
spin the water turbine. Conversely, in the proposed struc- the investigated time domain. Similar to the case of the
ture, as shown in Figure 7B, there are two turbine cycles 100 mL/min flow rate, the maximum magnitude of the
within the investigated time domain. One turbine cycle output voltage of the electromagnetic generator was mea-
occupied a time period of 195 seconds. The working time sured to be larger than 3 V at 200 mL/min.
in a turbine cycle was equal to 30 seconds. Within the Figure 9 shows the AC output voltage from the elec-
working time, the water turbine worked at high effi- tromagnetic generator in the traditional and proposed
ciency. Thus, a maximum AC output voltage greater than structures when the rainwater flow rate was 300 mL/
3 V was obtained from the electromagnetic generator. min. The average magnitude of the measured AC output
When the water flow rate was increased from voltage of the electromagnetic generator in the traditional
100 mL/min to 200 mL/min, the magnitude of the mea- structure could increase up to approximately 1 V in the
sured AC output voltage of the electromagnetic generator investigated time domain, as shown in Figure 9A. In
12 BAO AND WANG

F I G U R E 8 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
200 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is
200 mL/min

contrast, whilst using the proposed structure, there were increase when the rainwater flow rate exceeded 600 mL/
six turbine cycles in the investigated time domain. One min. Whilst using the proposed structure, 12 turbine
turbine cycle occupied approximately 65 seconds. The cycles existed in the investigated time domain at 600 mL/
working time within one turbine cycle was approxi- min. As shown in Figure 10B, one turbine cycle occupied
mately 25 seconds. The maximum magnitude of the out- approximately 34 seconds, and the working time in one
put voltage of the electromagnetic generator was turbine cycle consisted of 29 seconds. When the rainwa-
determined to be larger than 3 V, as shown in Figure 9B. ter flow rate increased above 900 mL/min, as shown in
Figures 10-12 show the AC output voltages from the Figure 11B and 12B, there was no waiting time in each
electromagnetic generator of the traditional and proposed turbine cycle. The water turbine worked efficiently in all
structures when the rainwater flow rates were 600, 900, the turbine cycles. Therefore, the proportion of working
and 1200 mL/min, respectively. It can be observed that time in the turbine cycle increased with the increase in
the average magnitude of the measured AC output volt- water flow rate owing to the decrease in the waiting time
age of the electromagnetic generator in the traditional in the turbine cycle, as shown in Figure 13. It can be
structure reaches approximately 2 V when the rainwater observed that the proportion of working time in the tur-
flow rate increased to 600 mL/min. However, the average bine cycle is 100% when the rainwater flow rate exceeds
amplitude of the measured AC output voltage did not 900 mL/min.
BAO AND WANG 13

F I G U R E 9 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
300 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is
300 mL/min

To further compare the energy harvesting perfor- 200 mL/min, as shown in Figure 8A, the corresponding
mance between the traditional structure and the pro- DC output charging voltage across the 680 μF capacitor is
posed structure under different flow rate conditions, the still 0 V, as shown in Figure 14B. This is because 0.5 V is
DC output charging voltage curves across the 680 μF smaller than the threshold voltage (0.6-0.7 V) of the
capacitor in the energy harvesting circuit connected to diodes that are used in the energy harvesting circuit.
the electromagnetic generator are illustrated in Therefore, the energy harvesting circuit in the traditional
Figure 14. As the water turbine in the traditional struc- structure cannot work at a flow rate of 200 mL/min. Con-
ture cannot work at 100 mL/min, as shown in Figure 7A, versely, the 680 μF capacitor was charged to 2.6 V after
the DC voltage across the 680 μF capacitor is equal to four turbine cycles using the proposed structure at
0 V, as shown in Figure 14A. In contrast, the 680 μF 200 mL/min in the same time domain. When the rainwa-
capacitor was charged to 2.4 V after two turbine cycles ter flow rate increased to 300 mL/min, as shown in
using the proposed structure at 100 mL/min for the same Figure 14C, the maximum DC charging voltage across
time duration. the 680 μF capacitor using the traditional and proposed
Although the average magnitude of the measured AC structures was measured to be 0.9 V and 2.7 V, respec-
output voltage of the electromagnetic generator in the tively. As the flow rate increased, the difference between
traditional structure reaches approximately 0.5 V at the maximum DC charging voltages across the 680 μF
14 BAO AND WANG

F I G U R E 1 0 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
600 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is 600 mL/min

F I G U R E 1 1 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
900 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is 900 mL/min

capacitor achieved using the traditional and proposed where C denotes the capacitance value of the capacitor,
methods became smaller, as shown in Figure 14D,F. as listed in Table 3. Ud denotes the maximum DC charg-
Table 3 lists the maximum DC charging voltages across ing voltage across the capacitor.
the 680 μF capacitor using the traditional and proposed Figure 15 shows the maximum energy stored in the
structures under different water flow rate conditions in 680 μF capacitor using the traditional and proposed
the same time domain. structures under different rainwater flow rates in the
The energy stored in the capacitor C can be calculated investigated time domain. The corresponding values of
as: the maximum energy stored in the 680 μF capacitor are
elaborately listed in Table 3. It can be observed that no
1 electrical energy was extracted and charged in the capaci-
E = CU d 2 , ð15Þ
2 tor of the energy harvesting circuit using the traditional
BAO AND WANG 15

F I G U R E 1 2 A, AC output voltage from the electromagnetic generator of the traditional structure when the rainwater flow rate is
1200 mL/min; B, AC output voltage from the electromagnetic generator of the proposed structure when the rainwater flow rate is
1200 mL/min

higher efficiency than the continuous power generation

of the traditional structure under different low water flow
rate conditions. The proposed structure can charge the
680 μF capacitor with nine times the electrical energy
stored using the traditional structure in the same amount
of time at 300 mL/min. However, this advantage dimin-
ishes as the flow rate increases. As shown in Figure 16,
the additional energy stored in the capacitor using the
proposed method, when compared to the traditional
method, decreases with the increase in the flow rate. At
1200 mL/min, as listed in Table 3, only 1.6 times the elec-
trical energy stored using the traditional structure was
charged in the 680 μF capacitor by using the proposed
structure for the same amount of time. In addition,
Figure 17 shows the results of a hygro-thermometer test
conducted on the proposed structure, which clearly
F I G U R E 1 3 Proportion of working time in the turbine cycle
shows the working mechanism and potential applications
of the proposed structure under different rainwater flow rate
of the proposed method.
conditions [Colour figure can be viewed at wileyonlinelibrary.com]

structure when the flow rate was below 200 mL/min. In 5 | CONCLUSION
contrast, the proposed structure was able to harvest elec-
trical energy below the ultra-low water flow rate; This paper introduced a pipe-connected tank consisting
e.g. 1958.4 μJ at 100 mL/min and 2298.4 μJ at 200 mL/ of a passive rainwater buffer incorporated into the tradi-
min. When the rainwater flow rate exceeded 200 mL/ tional rooftop rain energy harvester that uses water tur-
min, the traditional structure commenced continuous bines with electromagnetic generators for enhancing the
power generation. With the increase in the rainwater energy harvesting performance in light rain conditions.
flow rate, more electrical energy can be collected in the The pipe-connected tank was installed between the roof-
capacitor. However, as shown in Figure 15, the intermit- top gutters of the rainwater harvesting system and the
tent power generation of the proposed structure exhibits vertical rainwater pipe connected to the rainwater
16 BAO AND WANG

F I G U R E 1 4 DC output voltage from the energy harvesting circuit connected to the electromagnetic generator in the traditional and
proposed structures when the rainwater flow rate is (A) 100 mL/min; (B) 200 mL/min; (C) 300 mL/min; (D) 600 mL/min; (E) 900 mL/min;
(F) 1200 mL/min [Colour figure can be viewed at wileyonlinelibrary.com]

T A B L E 3 Maximum DC charging
Traditional structure Proposed structure
voltage and energy stored in the 680 μF
Flow rate (mL/min) DC voltage Stored energy DC voltage Stored energy capacitor using the traditional and
100 0V 0 μJ 2.4 V 1958.4 μJ proposed structures under different
water flow rate conditions
200 0V 0 μJ 2.6 V 2298.4 μJ
300 0.9 V 275.4 μJ 2.7 V 2478.6 μJ
600 1.5 V 765 μJ 2.75 V 2571.3 μJ
900 1.8 V 1101.6 μJ 2.8 V 2665.6 μJ
1200 2.2 V 1645.6 μJ 2.8 V 2665.6 μJ

harvesting tank. The working principle of the proposed energy stored using the conventional structure when the
structure and the mechanism of the passive rainwater rainwater flow rate was below 300 mL/min in the same
buffer were investigated and discussed in detail. A small- amount of time. However, this performance gap nar-
scale experimental prototype for the proposed structure rowed as the rainfall increased. Therefore, the advantage
was built for demonstrating the enhancement of energy of the proposed structure to enhance rain energy
harvesting performance in light rains when compared to harvesting performance is not significant under heavy
the traditional structure. The experimental results rain conditions. The rainwater buffer will rotate at a high
showed that the proposed structure extracted more rain speed under the impact of heavy rainfall and will be use-
energy than the traditional structure in light rain condi- less for enhancing the rain energy harvesting perfor-
tions. The proposed structure could charge a 680 μF mance as the rainwater flow rate is high in heavy rains.
capacitor with at least nine times the amount of electrical In addition, the buffer demonstrated no additional effects
BAO AND WANG 17

F I G U R E 1 5 Maximum energy stored in the capacitor using

the traditional and proposed structures under different rainwater
flow rates in the investigated time domain [Colour figure can be
viewed at wileyonlinelibrary.com]

F I G U R E 1 7 Hygro-thermometer powered by the proposed

structure [Colour figure can be viewed at wileyonlinelibrary.com]

F I G U R E 1 6 Additional energy stored in the 680 μF capacitor

using the proposed method in comparison with the traditional
method [Colour figure can be viewed at wileyonlinelibrary.com] ACKNOWLEDGEMENTS
This research is supported by Shenzhen Science and
Technology Planning Project (Grant
for energy harvesting in heavy rains. The pipe-connected JCYJ20170817111535167) and Key Scientific Research
tank will dynamically maintain a state where the tank is Projects of Universities in Guangdong Province (Grant
completely filled with rainwater during heavy rains; con- 2018KZDXM064). All the projects and funds are from
sequently, the vertical rainwater pipeline has sufficient China.
water pressure for energy harvesting using water turbines
with electromagnetic generators during the entire period ORCID
of heavy rainfall. In the future, we plan to design Bin Bao https://orcid.org/0000-0002-2315-3128
enhanced structures with more than one buffer to further
improve the proportion of working time in the turbine NOMENCLATURE
cycle and the energy harvesting efficiency of the electro- a one-half of the width of the inner wall of the
magnetic rain energy harvester in light rain square tank
environments. A rainwater block when 0 ≤ θ < tan−1([hs-am]/2a)
18 BAO AND WANG

am side length of the square section of the mass different stiffness levels for vibration energy harvesting. Smart
block Mater Struct. 2012;21(6):065017.
9. Zhang B, Ducharne B, Gupta B, Sebald G, Guyomar D, Gao J.
B rainwater block when tan−1([hs-am]/
−1 Experimental Sea wave energy extractor based on piezoelectric
2a) ≤ θ < tan (hs/2a) Ericsson cycles. J Intell Mater Syst Struct. 2018;29(6):1102-1112.
C rainwater block when tan−1(hs/ 10. Wu Y, Qiu J, Zhou S, Ji H, Chen Y, Li S. A piezoelectric spring
2a) ≤ θ < π/2-tan−1(am/[hs-am]) pendulum oscillator used for multi-directional and ultra-low
D rainwater block when π/2- tan−1(am/ frequency vibration energy harvesting. Appl Energ. 2018;231:
[hs-am]) ≤ θ < π/2 600-614.
E induced electromotive force (emf) 11. Viet NV, Xie XD, Liew KM, Banthia N, Wang Q. Energy
harvesting from ocean waves by a floating energy harvester.
hg height between the shaft and the bottom of
Energy. 2016;112:1219-1226.
the tank
12. Viet NV, Wang Q, Carpinteri A. Development of an ocean
hs height of the inner wall of the square tank wave energy harvester with a built-in frequency conversion
I magnitude of the current induced in the coil function. Int J Energ Res. 2018;42(2):684-695.
Mrw torque exerted by the rainwater in the buffer con- 13. Viet NV, Al-Qutayri M, Liew KM, et al. An octo-generator for
tainer on the shaft O energy harvesting based on the piezoelectric effect. Appl Ocean
N number of turns of a coil Res. 2017;64:128-134.
Ti thickness of the bottom of the square tank 14. Ong ZZ, Wong VK, Ho JH. Performance enhancement of a pie-
zoelectric rain energy harvester. Sens Actuators A: Phys. 2016;
Ts thickness of the wall of the square tank
252:154-164.
θ angle between the water surface and the plane of 15. Acciari G, Caruso M, Miceli R, et al. Piezoelectric rainfall
the opening of the container energy harvester performance by an advanced Arduino-based
ρc density of the square tank measuring system. IEEE Trans Ind Appl. 2017;54(1):458-468.
ρm equivalent density of the mass block 16. Rajeev SP, John SK, Cherian R, et al. Next-generation rooftop
ρwater density of rainwater tribo–piezo electric energy harvesting from rain power. Appl
Nanosci. 2020;10(3):679-686.
ORCID 17. Wong CH, Dahari Z, Manaf AA, et al. Harvesting raindrop
energy with piezoelectrics: a review. J Electron Mater. 2015;44
Bin Bao https://orcid.org/0000-0002-2315-3128
(1):13-21.
18. Jellard SCJ, Pu SH, Chen S, Yao K, White NM. Water droplet
R EF E RE N C E S impact energy harvesting with P (VDF-TrFE) piezoelectric can-
1. Shrivastava PK, Patel DP, Nayak D, Satasiya KF, Patel DC. tilevers on stainless steel substrates. Smart Mater Struct. 2019;
Harvesting and potable use of rooftop rain water to tackle 28(9):095002.
imminent drinking water crisis in coastal Gujarat, India. Curr J 19. Wang Y, Duan J, Zhao Y, Yuan H, He B, Tang Q. Film-type
Appl Sci Technol. 2019;35(6):1-10. rain energy converters from conductive polymer/PtCo hybrids.
2. Abdulla F. Rainwater harvesting in Jordan: potential water sav- Appl Energ. 2018;218:317-324.
ing, optimal tank sizing and economic analysis. Urban Water J. 20. Bao B, Chen W, Wang Q. A piezoelectric hydro-energy har-
2019;1(1):1-11. vester featuring a special container structure. Energy. 2019;189:
3. Joleha J, Mulyadi A, Wawan W, et al. Application of rainwater 116261.
harvesting technology to supply sustainable domestic water. Int 21. Bullen CR. A low energy housing design in an area of high
J Elec Energ Power Syst Eng. 2019;2(1):11-16. wind and rain: An Innovative Housing Scheme at Stenness on
4. Siddiqui R, Siddiqui S. Assessing the rooftop rainwater the Orkney Islands. Energ Build. 2000;32(3):319-326.
harvesting potential in urban residential areas of Pakistan: a 22. Xie XD, Wu N, Yuen KV, Wang Q. Energy harvesting from
case study of model town, Lahore, Pakistan. Int J Econ Environ high-rise buildings by a piezoelectric coupled cantilever with a
Geol. 2019;9(2):11-19. proof mass. Int J Eng Sci. 2013;72:98-106.
5. Rahmat SN, Al-Gheethi AAS, Ayob S, et al. Development of 23. Edeoja AO, Ekoja M, Tuleun LT. Effect of penstock area reduc-
dual water supply using rooftop rainwater harvesting and tion and number of turbine V-blades on the performance of a
groundwater systems. SN Appl Sci. 2020;2(1):85-93. simple Pico-hydropower system. Eur J Adv Eng Technol. 2017;4
6. Acciari G, Caruso M, Fricano M, et al. Experimental investi- (11):797-806.
gation on different rainfall energy harvesting structures. 24. Zhou D, Deng ZD. Ultra-low-head hydroelectric technology: a
2018 Thirteenth International Conference on Ecological Vehi- review. Renew Sustain Energy Rev. 2017;78:23-30.
cles and Renewable Energies (EVER). Monte-Carlo: IEEE; 25. Fan K, Cai M, Wang F, et al. A string-suspended and driven
2018:1-5. rotor for efficient ultra-low frequency mechanical energy
7. Guigon R, Chaillout JJ, Jager T, Despesse G. Harvesting rain- harvesting. Energ Convers Manage. 2019;198:111820.
drop energy: experimental study. Smart Mater Struct. 2008;17 26. Lallart M, Lombardi G. Synchronized switch harvesting on
(1):015039. ElectroMagnetic system: a nonlinear technique for hybrid
8. Xu JW, Liu YB, Shao WW, Feng Z. Optimization of a right- energy harvesting based on active inductance. Energ Convers
angle piezoelectric cantilever using auxiliary beams with Manage. 2020;203:112135.
BAO AND WANG 19

27. Juneau TL. Combined water tank and generator: U.S. Patent 33. MMSRS B, Anbuudayasankar SP, Balaji K. Power generation
8,080,892. December 20, 2011. by high head water in a building using micro hydro turbine—
28. Yi JH, Cho YC, Lin CT. Method and apparatus for supplying a greener approach. Environ Sci Pollut R. 2016;23(10):9381-
fresh water and generating electricity by collected rainwater: 9390.
U.S. Patent 7,348,685. March 25, 2008. 34. Lallart M, Priya S, Bressers S, et al. Small-scale piezoelectric
29. Desai A, Mukhopadhyay I, Ray A. Theoretical analysis of a energy harvesting devices using low-energy-density sources.
Pico-hydro power system for energy generation in rural or J Korean Phys Soc. 2010;57(41):947-951.
isolated area. 2014 IEEE PES Asia-Pacific Power and Energy 35. Yunfeng G. The principle and design of the QiQi container
Engineering Conference (APPEEC). Hong Kong, China: device. Mech Eng. 1999;21(2):76-79.
IEEE; 2014:1-4.
30. Kumar KR, Kulgod SP, Surendran A. Simulation and analysis of
energy harvesting from grey water and rain water in high rises.
2016 IEEE Region 10 Symposium (TENSYMP). Bali, Indonesia: How to cite this article: Bao B, Wang Q. Small-
IEEE; 2016:404-409.
scale experimental study on the optimisation of a
31. Muhsen H, Ibrahim M, Alsheikh A, et al. Turbine design and
its impact on energy harvesting from in-pipe hydro systems. Int
rooftop rainwater energy harvester using
J Mech Eng Robot Res. 2019;8(5):685-690. electromagnetic generators in light rains. Int
32. Martin S, Sharma AK. Analysis on rainwater harvesting and its J Energy Res. 2020;1–19. https://doi.org/10.1002/
utilization for pico hydro power generation. Int J Adv Res Com- er.5726
put Eng Technol. 2014;3:2121-2126.