sensors Article
A Self-Powered, Threshold-Based Wireless Sensor for the Detection of Floor Vibrations Byung C. Jung *, Young Cheol Huh and Jin-Woo Park Department of System Dynamics, Korea Institute of Machinery and Materials, Daejeon 34103, Korea;
[email protected] (Y.C.H.);
[email protected] (J.-W.P.) * Correspondence:
[email protected]; Tel.: +82-42-868-7463 Received: 15 October 2018; Accepted: 3 December 2018; Published: 5 December 2018
Abstract: Smart buildings will soon be a reality due to innovative Internet of Things (IoT) applications. IoT applications can be employed not only for energy management in a building, but also for solving emerging social issues, such as inter-floor noise-related disputes in apartments and the solitary death of an elderly person. For example, acceleration sensors can be used to detect abnormal floor vibrations, such as large vibrations due to jumping children or unusual vibrations in a house where an elderly person is living alone. However, the installation of a conventional accelerometer can be restricted because of the sense of privacy invasion. In this study, a self-powered wireless sensor using a threshold-based method is studied for the detection of floor vibrations. Vibration levels of a bare slab in a testbed are first measured when a slab is impacted by a bang machine and an impact ball. Second, a piezoelectric energy harvester using slab vibration is manufactured to generate electrical power over a threshold. Next, the correlation among harvested energy, floor vibration, and impact noise is studied to check whether harvested energy can be employed as a condition detection threshold. Finally, a prototype of a self-powered wireless sensor to detect abnormal conditions in floor vibrations is developed and its applicability is demonstrated. Keywords: vibration-based energy harvester; piezoelectric energy harvester; detection of floor vibration; smart building
1. Introduction For the successful realization of a Smart Building [1,2], connected wireless sensors are required to detect abnormal changes in a building and make smart decisions for residents. In recent years, Internet of Things (IoT) applications with energy management technologies for building automation [3,4] have been widely developed. In addition, with the increasing urbanization and individualism of a society, studies on monitoring systems to resolve social issues, such as noise-related disputes in apartments and the solitary death of an elderly person living alone, have also been performed [5–9]. Lam et al. [5] developed an occupant detection method by monitoring and analyzing step-induced structural vibration in an apartment. Mun et al. [6] researched the relationships among impact force, the acceleration response of a slab, and floor impact noise. Acceleration sensors are essential to those applications; however, in some cases, privacy invasion and cyber security problems can restrict the usage of conventional sensors. For example, floor noise-related disputes in apartments have increased over the past few years and became a crucial social issue in Asian countries. Monitoring of acceleration levels with conventional accelerometers installed on floors of an apartment can solve this to support a mediation process by providing objective data on human-induced floor vibrations. However, the realization of it may not be feasible since vibration signals being measured in real-time during the day and night can be seen as privacy invasion. For example, residence patterns analyzed
Sensors 2018, 18, 4276; doi:10.3390/s18124276
www.mdpi.com/journal/sensors
Sensors 2018, 18,18, x FOR Sensors 2018, 4276 PEER REVIEW
2 of2 of 1212
analyzed with real-time signals can be used in crime investigations, especially when a monitoring with real-time signals be usedsystem. in crime investigations, especially when a monitoring system is system is developed withcan a wireless developed with a wireless system. Energy harvesting that converts ambient, otherwise wasted, energy sources into usable electricity Energy harvesting thatasconverts ambient, otherwise wasted,sensing energy sources into usable has been widely researched a promising solution for wireless nodes. Among the electricity various has been widely researched as a promising solution for wireless sensing nodes. Among the various energy harvesting devices, a piezoelectric energy harvester that generates electricity from mechanical energy harvesting devices, a piezoelectric energy harvester that generates electricity from mechanical strain energy has received much attention as a power source of wireless sensors due to its high power strain and energy much attention power source of wireless sensorsby due its high density easehas of received installation [10–13]. Whenasa apiezoelectric material is deformed antoexternal power density and ease of installation [10–13]. When a piezoelectric material is deformed by an mechanical force, it generates a charge flow due to the change in electric polarization. This phenomenon mechanical force, it generates due to the change in electric This is external called the piezoelectric effect [14,15]. Aa charge typical flow piezoelectric energy harvester is apolarization. cantilever-type phenomenon is called the piezoelectric effect [14,15]. A typical piezoelectric energy harvester unimorph/bimorph using a ‘31’ mode (where mechanical strain is perpendicular to the poling direction). is a cantilever-type unimorph/bimorph using a ‘31’ mode (where mechanical strain is perpendicular This generates AC voltage proportional to the bending strains of a piezoelectric material [16–19], and isto the polingtodirection). generates AC voltage proportional to the bending strains of a piezoelectric connected interface This electrical circuits designed by an impedance matching theory to extract material [16–19], and is connected to interface electrical circuits designed by an impedance matching maximum harvesting power [20–22]. theory to extract maximum harvesting power [20–22]. In this paper, a self-powered, threshold-based wireless sensor is proposed to detect abnormal In this paper, a self-powered, threshold-based wireless sensor is proposed to detect abnormal floor vibration conditions. A cantilever-type piezoelectric energy harvester is employed to generate floor vibration conditions. A cantilever-type piezoelectric energy harvester is employed to generate electric energy proportional to the amount of mechanical floor vibration. One benefit of employing proportional to the amount of mechanical floor vibration. Onesource benefitfor ofaemploying anelectric energyenergy harvester is that accumulated electric energy can be used as a power wireless an energy harvester is that accumulated electric energy can be used as a power source for a wireless data transmitter. Figure 1 shows the concept of the proposed self-powered wireless sensor. When data transmitter. Figure 1 shows the concept of the proposed self-powered wireless sensor. When is the the vibration of a slab is below a designated threshold, the amount of harvested electric energy vibration of a slab below aadesignated threshold, the amount of harvested electric is energy is not high not high enough toisoperate wireless transmitter; thus, the wireless transmitter not activated. enough to operate a wireless transmitter; thus, the wireless transmitter is not activated. However, when However, when the vibration of a slab is high enough to generate electric energy above that the vibration of a slab transmitter is high enough to generate energy threshold, the wireless threshold, the wireless is activated andelectric transmits an above alarm that signal to a receiver. The transmitter is activated and transmits an alarm signal to a receiver. The proposed self-powered proposed self-powered wireless sensor has the following advantages: (1) It is relatively free from wirelessofsensor has the following (1) Itsave is relatively free from invasion of privacy, since invasion privacy, since it does advantages: not collect and vibration signals in real-time, and only it doesdata not collect save vibration signals real-time, and only collects duringchemical abnormal collects duringand abnormal conditions; (2) in there is no maintenance cost data to change conditions; (2) there is no maintenance cost to change chemical batteries; and (3) installation is easy and batteries; and (3) installation is easy and inexpensive compared to wired systems that require inexpensive compared to wired systems that require wiring for data transmission and power supply. wiring for data transmission and power supply.
(a) Normal condition
(b) Abnormal condition
Figure 1. Concept of the self-powered wireless sensor forfor threshold-based Figure 1. Concept of the self-powered wireless sensor threshold-baseddetection detectionofoffloor floorvibrations. vibrations.
The rest this article organized follows. Section 2 provides a description of developing The rest ofof this article is is organized as as follows. Section 2 provides a description of developing a a cantilever-type piezoelectric energy harvester designed based on the dynamic characteristics cantilever-type piezoelectric energy harvester designed based on the dynamic characteristics of aof a slab a testbed. Section 3 shows correlation studies among slab vibrations, floor impact slab in a in testbed. Section 3 shows correlation studies among slab vibrations, floor impact sound,sound, and and harvested energy. A strong correlation among is an essential requirement harvested electricelectric energy. A strong correlation among these these factorsfactors is an essential requirement for indirectly detecting floor vibrations and the corresponding impact noise with harvested electric
Sensors 2018, 18, 18, 4276 x FOR PEER REVIEW Sensors 2018,
33of of 12 12
energy. Section 4 describes the integration of the energy harvester with a wireless transmission for indirectly floor vibrationsofand theself-powered corresponding impact sensor noise with electric system, and detecting gives a demonstration the wireless for harvested the detection of energy. Section 4 describes the integration of the energy harvester with a wireless transmission system, floor vibrations. and gives a demonstration of the self-powered wireless sensor for the detection of floor vibrations. 2. Development of a Cantilever-Type Piezoelectric Energy Harvester 2. Development of a Cantilever-Type Piezoelectric Energy Harvester 2.1. Testbed and Dynamic Characteristics of the Slab 2.1. Testbed and Dynamic Characteristics of the Slab Before designing a piezoelectric energy harvester, the dynamic characteristics of the slab in the Before designing a piezoelectric energy harvester, the dynamic characteristics of the slab in the testbed are experimentally analyzed. The testbed consists of a reverberation room and a concrete slab testbed are experimentally analyzed. The testbed consists of a reverberation room and a concrete slab located on top of the reverberation room. The shape and size of the reverberation room are shown located on top of the reverberation room. The shape and size of the reverberation room are shown in in Figure 2a. Figure 2b shows the reverberation room for measuring impact noise. Figure 2c,d show Figure 2a. Figure 2b shows the reverberation room for measuring impact noise. Figure 2c,d show the the concrete slab, and the thickness of the slab is 150 mm. concrete slab, and the thickness of the slab is 150 mm.
(a) Testbed arrangement
(b) Reverberation room
(c) Setup of bang machine tests
(d) Setup of impact ball tests
Figure and test test setup. setup. Figure 2. 2. Testbed Testbed configuration configuration and
Standard heavyweight machine andand an impact ball)ball) that that are officially used Standard heavyweight impact impactsources sources(a(abang bang machine an impact are officially for measuring floor impact soundsound are employed to excite concrete slab. Figure 2c,d show setups used for measuring floor impact are employed to the excite the concrete slab. Figure 2c,dthe show the of the bang machine and impact ball tests. While impacting the center of the bare slab with the bang setups of the bang machine and impact ball tests. While impacting the center of the bare slab with the machine and the ball, vibration signals are measured withwith an accelerometer attached on the bang machine andimpact the impact ball, vibration signals are measured an accelerometer attached on center of the FigureFigure 3a–d show measured acceleration signals in time and frequency domains. the center ofslab. the slab. 3a‒d the show the measured acceleration signals in time and frequency As shown in Figure 3b,d, the natural frequency of the slab is 24.8 Hz. During the bang machine domains. 2 and test, As theshown maximum amplitude andnatural time duration of of thethe acceleration are about 4.0 m/s in Figure 3b,d, the frequency slab is 24.8signal Hz. During the bang machine 0.8 s,the respectively. regard and to thetime impact ball test, the acceleration maximum amplitude time4.0 duration of test, maximumWith amplitude duration of the signal areand about m/s2 and 2 the s, acceleration signal aboutto1.55 andball 0.6test, s, respectively. Theamplitude impact sources and dynamic 0.8 respectively. Withare regard them/s impact the maximum and time duration 2 and characteristics of the signal slab areare summarized Table 1. 0.6 s, respectively. The impact sources and of the acceleration about 1.55inm/s dynamic characteristics of the slab are summarized in Table 1.
Sensors 2018, 18, 4276
4 of 12
Sensors 2018, 18, x FOR PEER REVIEW
4 of 12
(a) Time domain signal (Bang machine test)
(b) Frequency domain signal (Bang machine Test)
(c) Time domain signal (Impact ball test)
(d) Frequency domain signal (Impact ball test)
Figure3.3.Measured Measured acceleration Figure accelerationsignals. signals. Table Dynamic characteristics characteristics ofofthe slab. Table 1.1.Dynamic thebare bare slab. Impact Source Impact Source Bang machine Bang machine Drop height 0.85 m Drop height 0.85 mkPa Tire pressure 240 Tire pressure kPa N Impact force (peak) 240≈4000 ImpactTime forceduration (peak) ≈4000 ≈20Nms Time duration ≈ 20 ms Impact ball Impact ball Drop height 1m Impact force (peak) ≈1500 Drop height 1m N Time duration ≈20Nms Impact force (peak) ≈1500
Natural Frequency Acceleration Acceleration Duration of of Natural Frequency (PeakTime Time Duration Slab (Peak to Peak) SlabSlab Vibration of of Slab to Peak) Vibration 24.8 Hz 24.8 Hz
24.8 Hz
24.8 Hz
−3.0 m/s2
2 −3.0 m/s ~4.0 m/s2 ~4.0 m/s2
−1.17 m/s2
2 2 ~1.55 − 1.17m/s m/s ~1.55 m/s2
≈0.8 s
≈0.8 s
≈0.6 s
≈0.6 s
Time duration ≈20 ms 2.2. Design of the Cantilever-Type Piezoelectric Energy Harvester
A of piezoelectric energy harvester is designed produce the required energy (threshold) at a 2.2. Design the Cantilever-Type Piezoelectric Energy to Harvester
given impact condition. For wireless data transmission, this study employs data A piezoelectric energy harvester is designed toA) produce the required Heuvelton, energy (threshold) a given transmitting/receiving nodes, AmbioMote24 (Type from Ambiosystems, NY, USA at [23]. impact condition. For wireless data transmission, this study employs data transmitting/receiving The data transmitting/receiving nodes require 5.3 V for initiating, 2.7 V for data transmitting, and nodes, AmbioMote24 (Type A) Ambiosystems, Heuvelton, NY, USA [23]. The about 15 μW for continuous datafrom monitoring [10]. By considering the first resonance frequency of data the slab and the power generation capacity tens of microwatt fordata wireless data transmission), transmitting/receiving nodes require 5.3 V(several for initiating, 2.7 V for transmitting, and about a piezoelectric bimorph specimen from Piezo Inc., Cambridge, MA,resonance USA [24] of a 63.5 mmof the 15 µW for continuous data monitoring [10]. System By considering the first frequency and a 31.8 generation mm width iscapacity selected. Figure 4 and Table 2 show the shape and specification of the slab length and the power (several tens of microwatt for wireless data transmission), base bimorph specimen. The specimen consists of two piezoelectric plates with nickel electrode a piezoelectric bimorph specimen from Piezo System Inc., Cambridge, MA, USA [24] of a 63.5 mm layers, a center brass shim, and two adhesive layers.
length and a 31.8 mm width is selected. Figure 4 and Table 2 show the shape and specification of the base bimorph specimen. The specimen consists of two piezoelectric plates with nickel electrode layers, a center brass shim, and two adhesive layers.
Sensors 2018, 18, 4276 Sensors 2018, x FOR PEER REVIEW Sensors 2018, 18,18, x FOR PEER REVIEW
5 of 12 5 of 12 12 5 of
Figure 4. Shape and size of the base bimorph specimen. Figure4.4.Shape Shapeand andsize sizeof ofthe thebase basebimorph bimorphspecimen. specimen. Figure Table 2. Specification of the bimorph specimen [24]. Table 2. Specification of the bimorph specimen [24]. Table 2. Specification of the bimorph specimen [24].
Specification Value Value Value Model number D220-A4-503YB Model number D220-A4-503YB Model number D220-A4-503YB Piezo material PZT-5A Piezo material PZT-5A material SizePiezo of a base specimen 63.5 PZT-5A × 31.8 mm Size of a base specimen 63.5 × 31.8 mm Active length of the bimorph (Figure 5a) 57.15 mmmm Size of a base specimen 63.5 31.8 Active length of the bimorph (Figure 5a) 57.15× mm Weight 10.4 gram Active length ofWeight the bimorph (Figure 5a) 57.15 mm 10.4 gram PZT thickness 0.019 mm PZT thickness 0.019 mm Weight 10.4 gram Brass thickness 0.013 mm Brass thickness 0.013 mm PZT thickness 0.019 mm Capacitance 232 µF Capacitance 232 μF Brass thickness 0.013 mm Natural frequency (without mass) 52 Hz Natural frequency (without mass) 52 Hz Capacitance 232 μF Specification Specification
Natural frequency (without mass)
(a) Bimorph and electric circuit for fluency tuning
(a) Bimorph and electric circuit for fluency tuning
52 Hz
(b) Tuned bimorph
(b) Tuned bimorph
(c) Test facilities Figure5.5.Test Testsetup setupfor forfrequency frequency tuning. tuning. Figure (c) Test facilities
Next, base bimorph specimenininFigure Figure44isistuned tuned to to match match aa resonance Next, thethe base bimorph specimen resonancefrequency frequencywith withthe the Figure 5. Test setup for frequency tuning. natural frequency of the Figure 5 shows the configuration thesetup test setup for frequency natural frequency of the slab.slab. Figure 5 shows the configuration of theoftest for frequency tuning. The base is specimen to (ET-139, a shaker (ET-139, Labworks Inc., Novato, CA, using USA) using a Thetuning. base specimen fixed toisa fixed shaker Labworks Inc., Novato, CA, USA) a testing Next, the base bimorph specimen in Figure 4 is tuned to match a resonance frequency with the testing fixture. shaker is controlled with a controller (M + P Vibcontrol, m + p international, Inc., fixture. The shakerThe is controlled with a controller (M + P Vibcontrol, m + p international, Inc., Verona, natural frequency ofand theanslab. Figure 5 shows theDytran configuration of the test setupCA, for USA) frequency Verona, NJ, USA) accelerometer (3055B2T, Instruments, Chatsworth, NJ, USA) and an accelerometer (3055B2T, Dytran Instruments, Chatsworth, CA, USA) in Figurein5c. tuning. base specimenofisa fixed to alength shaker Labworks Inc., as Novato, CA, USA) using FigureThe 5c. A plastic 12.7ismm is (ET-139, used to bond tip masses, shown5a. in Figure For a A plastic plate of a 12.7plate mm length used to bond tip masses, as shown in Figure For the5a. voltage testing fixture. The shaker is controlled with a controller (M + P Vibcontrol, m + p international, Inc., the voltage monitoring, from are the connected specimen are to a resistance resistor. The monitoring, two wires fromtwo thewires specimen to aconnected resistor. The of resistance a resistor of (R)a is Verona, and accelerometer (3055B2T, Dytran Chatsworth, USA) in resistorNJ, (R)USA) is set to 10 an MΩ. Sine sweep vibration tests with aInstruments, 0.01 g acceleration peak areCA, performed set to 10 MΩ. Sine sweep vibration tests with a 0.01 g acceleration peak are performed by increasing Figure 5c. A plastic plate of a 12.7 mm length is used to bond tip masses, as shown in Figure by increasing tip masses while measuring the voltage (V) on a resistor. The total weight of the5a.tipFor tip masses while measuring the voltage (V) on a resistor. The total weight of the tip masses to make themasses voltage two wiresfrequency from theof specimen are connected a resistor. resistance of a to monitoring, make the first natural the bimorph 24.8 Hz (thetofirst natural The frequency of the the first natural frequency of the bimorph 24.8 Hz (the first natural frequency of the slab) is 15.14 g. resistor (R) is set to 10 MΩ. Sine sweep vibration tests with a 0.01 g acceleration peak are performed by increasing tip masses while measuring the voltage (V) on a resistor. The total weight of the tip masses to make the first natural frequency of the bimorph 24.8 Hz (the first natural frequency of the
Sensors 2018, 18, 4276 Sensors 2018, 18, x FOR PEER REVIEW
6 of 12 6 of 12
Sensors 2018, 18, x FOR PEER REVIEW 6 of 12 Figure shows the tuned bimorph, and Figure 6 shows the sine sweepthe vibration test result. At test the slab) is 5b 15.14 g. Figure 5b shows the tuned bimorph, and Figure 6 shows sine sweep vibration resonant frequency, thefrequency, maximumthe voltage of the voltage tuned bimorph is 2.94 V. result. At the resonant maximum of the tuned bimorph is 2.94 V. slab) is 15.14 g. Figure 5b shows the tuned bimorph, and Figure 6 shows the sine sweep vibration test
result. At the resonant frequency, the maximum voltage of the tuned bimorph is 2.94 V.
Figure test results the bimorph. Figure 6. Sine sweeptest testresults resultsofof of the bimorph. Figure6. 6. Sine Sine sweep sweep the bimorph.
3. Correlation Study 3. Correlation Study 3. Correlation Study The correlation among floor impact sound, slab vibrations, and harvested electric energy is TheThe correlation sound, slab slabvibrations, vibrations,and and harvested electric energy correlationamong amongfloor floor impact impact sound, harvested electric energy is is studied to confirm whether harvested electric energy cancan be used as aasquantitative threshold to predict studied confirmwhether whetherharvested harvested electric electric energy a quantitative threshold to to studied to to confirm energy canbebeused used as a quantitative threshold floor vibrations and the and corresponding impact sound in an apartment. As in Figure Figure7a, 7a, an predict floor vibrations thecorresponding corresponding impact anan apartment. Asshown shown in predict floor vibrations and the impactsound soundinin apartment. As shown in Figure 7a, accelerometer (Type 4370, B&K, Nærum, Denmark) is attached on the centerof of theconcrete concrete slab to an accelerometer (Type 4370, B&K, Nærum, Denmark) onon thethe center slab an accelerometer (Type 4370, B&K, Nærum, Denmark)isisattached attached centerthe of the concrete slab measure the floor acceleration of the slab. TheThe sound pressure level (SPL) in in thethe reverberation room to measure the floor acceleration of the slab. sound pressure level (SPL) reverberation to measure the floor acceleration of the slab. The sound pressure level (SPL) in the reverberation is also measured using five microphones (Tpye 4942-A-021, B&K, Nærum, Denmark), as shown room is also measured using five microphones B&K, Nærum, Denmark), as in room is also measured using five microphones (Tpye (Tpye4942-A-021, 4942-A-021, B&K, Nærum, Denmark), as shown As a data acquisition used SIRIUS Mini (DEWESoft, Figure 7a. in AsFigure a data7a. acquisition system, thissystem, study this usedstudy SIRIUS Mini (DEWESoft, Trbovlje,Trbovlje, Slovenia) to shown in Figure 7a. As a data acquisition system, this study used SIRIUS Mini (DEWESoft, Trbovlje, Slovenia) to measure voltagesand at capacitors and acceleration 3560C (B&K, Nærum, measure voltages at capacitors acceleration levels. PULSElevels. 3560CPULSE (B&K, Nærum, Denmark) is also Slovenia) to measure voltages at capacitors and acceleration levels. PULSE 3560C (B&K, Nærum, Denmark) is also used to measure sound pressure levels. Five locations shown in Figure 7b with used to measure sound pressure levels. Five locations shown in Figure 7b with circled numbers are Denmark) is also used to measure sound pressureand levels. Five locations shown Thus, in Figure 7bof with circled with numbers are machine impacted withan a bang machine an impact ball, a sequentially. a total impacted a bang and impact ball, sequentially. Thus, total of 10 tests are performed circled numbers are impacted with a bang machine and an impact ball, sequentially. Thus, a total of tests are performed measured signalsNo. are analyzed. Location No. ① 27 cm of away theavoid 1 is 27 cm and10measured signals areand analyzed. Location away from theiscenter the from slab to
10 tests are performed and interference measured signals are analyzed. Location No. ① is 27 cm away from the center of the slab to avoid with the accelerometer. interference with the accelerometer. center of the slab to avoid interference with the accelerometer.
(a) Test configuration
(a) Test configuration
(c) Locations of sensor and DAQ system
(b) Slab size and impact location
(b) Slab size and impact location
(d) Installation of accelerometer and bimorphs
Figure 7. Test setup for correlation study. (c) Locations of sensor and DAQ system (d) Installation of accelerometer and bimorphs
Figure 7. Test setup for correlation study.
Sensors 2018, 18, 4276 Sensors 2018, 18, x FOR PEER REVIEW
7 of 12 7 of 12
Figure 88 shows shows plots plots of of slab slab acceleration acceleration and and sound sound pressure pressure level, level, when when location location No. No. ① is 1 is Figure
impacted. The frequencies of of one-third one-third octave octave bands, bands, and and the the Y-axis Y-axis impacted. The X-axis X-axis refers refers to to the the center center frequencies −3 cm/s indicates accelerations accelerationsand andsound sound pressure levels in decibels. reference values 10−3 2cm/s indicates pressure levels in decibels. The The reference values are 10are and2 and 20 for μPaacceleration for acceleration and sound pressure, respectively. clear correlation at the first 20 µPa and sound pressure, respectively. A clearAcorrelation is foundisatfound the first natural natural frequency of (the Hz).for Except some frequencies soundispressure is frequency of the slab ≈25slab Hz).(≈25 Except some for frequencies where thewhere soundthe pressure amplified amplified due to room the correlation between slab acceleration soundlevels pressure levels is due to room modes, themodes, correlation between slab acceleration and soundand pressure is generally generally maintained. acceleration and SPL are summarized maintained. MeasuredMeasured acceleration and SPL levels are levels summarized in Table 3.in Table 3.
(a) Bang machine test
(b) Impact ball test
Figure 8. 8. Measured Measured acceleration acceleration and and sound sound pressure pressure levels. levels. Figure Table Table 3. 3. Summary Summary of of test test results. results. Impact Impact Machine Machine
Impact Impact Location Location
Distance from Distance from Accelerometer Accelerometer & & Bimorph Bimorph
Harvested Harvested Energy Energy (Average (Average of of Two Two Bimorphs) Bimorphs) (µJ) (µJ)
Square SquareRoot Root ofofHarvested Harvested Energy Energy(dB, (dB, Ref.:1 Ref.:1µJ)) µJ))
Acceleration Acceleration at 25 25 Hz Hz at (dB, Ref.: Ref.: (dB, −3 cm/s2 ) 10 2) 10−3 cm/s
SPL SPL at 25 at Hz 25Ref.: Hz (dB, (dB, 20 Ref.: 20 µPa) µPa)
Impact ImpactBall Ball Bang Bang Machine Machine Impact ImpactBall Ball Bang Bang Machine Machine Impact Ball Impact Ball Bang Bang Machine Machine Impact Ball Impact Bang Ball Bang Machine Machine Impact Ball Impact Bang Ball Bang Machine Machine
1 ①
27 27 cm cm
10.75 10.75
10.31 10.31
83.50 83.50
78.13 78.13
1
①
27 cm
78.11 78.11
18.93 18.93
91.29 91.29
85.62 85.62
1
①
151.7 cm 151.7 cm
0.95 0.95
−−0.22 0.22
75.46 75.46
69.77 69.77
1
① 1
①
151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm 151.7 cm
17.29 17.29 1.69 1.69 14.83 14.83 1.28 1.28 17.20 17.20 1.82 1.82 16.77 16.77
12.38 12.38 2.27 2.27 11.71 11.71 1.09 1.09 12.36 12.36 2.60 2.60 12.25 12.25
85.59 85.59 77.23 77.23 86.36 86.36 75.65 75.65 84.87 84.87 76.43 76.43 86.05 86.05
79.30 79.30 71.73 71.73 79.88 79.88 72.70 72.70 81.25 81.25 73.12 73.12 80.65 80.65
1
① 1
① 1
① 1
① 1
①
Next, bonded on the center of theofslab employed to measure the amount Next, two twotuned tunedbimorphs bimorphs bonded on the center theare slab are employed to measure the of harvested electric energy from slab vibrations in Figure 7d. Two tuned bimorphs are separately amount of harvested electric energy from slab vibrations in Figure 7d. Two tuned bimorphs are connected electric energy storage circuits. this study, Standard Energy Energy Harvesting (SEH) separately to connected to electric energy storage In circuits. In thisthe study, the Standard Harvesting interface circuit in Figure 9 is employed. Two wires from each bimorph are linked to a bridge rectifier (SEH) interface circuit in Figure 9 is employed. Two wires from each bimorph are linked to a bridge and a capacitor. While impacting the slab with bang machine impactand ball,impact voltageball, changes at rectifier and a capacitor. While impacting the aslab with a bangand machine voltage capacitors measured.are The harvestedThe energy of one bimorph is calculated usingisEquation (1):using changes atare capacitors measured. harvested energy of one bimorph calculated Equation (1):
Harvested Energy = 0.5·C·(V2 − V0 2 ), (1) Harvested Energy = 0.5·C·(V2 – V02), (1) where V0 and V refer to the measured voltage before and after excitation, respectively; and C refers where V0 and V refer to the measured voltage before and after excitation, respectively; and C refers to the capacitance of the capacitor. The capacitance (C) of the capacitor is set to 47 µF in this study. to the capacitance of the capacitor. The capacitance (C) of the capacitor is set to 47 μF in this study. The calculated harvested energies of the two bimorphs are finally averaged, as shown in the third
Sensors 2018, 18, 4276
8 of 12
Sensors 2018, x FORPEER PEERREVIEW REVIEW Sensors 2018, 18,18, x FOR
8 of8 12 of 12
The calculated harvested energies of the two bimorphs are finally averaged, as shown in the third column Table knownthat thataaaharvestable harvestablepower power (P) ofofaaacantilever-type cantilever-type energy column inin Table known that harvestable power(P) (P)of cantilever-typepiezoelectric piezoelectric energy column in Table 3.3.3. ItItIt isisis known piezoelectric energy harvester is proportional to the square of acceleration amplitude (A in ). Roundy et al. [10] developed an an harvester is proportional to the square of acceleration amplitude (A in ). Roundy et al. [10] developed harvester is proportional to the square of acceleration amplitude (Ain ). Roundy et al. [10] developed an analyticalequation equation for aa two-layer two-layer bimorph bimorph to predict the power based analytical the harvestable harvestable basedon on analytical equation for for a two-layer bimorph to predicttothepredict harvestable power based onpower one-dimensional one-dimensional beam theory, and prove the relationship (P ∝ A in2).2 Thus, the square root of one-dimensional beam the theory, and prove (P ∝ root Ain ).ofThus, the square root of beam theory, and prove relationship (P ∝ the Ain 2relationship ). Thus, the square harvested electric energy harvested electricenergy energyisisalso alsocalculated, calculated, as as shown in the fourth column ininTable 3. 3. harvested electric shown in the fourth column Table is also calculated, as shown in the fourth column in Table 3.
Figure 9. Configuration of electric circuits. Figure Figure9. 9. Configuration Configuration of of electric electriccircuits. circuits.
Figure 10 shows the relationships among the data in Table 3. The X-axis refers to acceleration Figure 10 shows relationships among the in Table X-axis refers acceleration shows the the relationships among the data data Table3. 3. The The refers to to acceleration andFigure sound10pressure levels, and the Y-axis refers to theinsquare root of X-axis the harvested energy and and sound pressure levels, and the Y-axis refers to the square root of the harvested energy and sound and sound pressure levels, the are Y-axis refers tointhe square As rootshown of theinharvested and sound pressure levels. All and values represented decibels. the plots,energy data on pressure levels.sound All values represented in decibels. shownto inbe the plots,indata acceleration, sound pressure levels. Allare values are represented in As decibels. As shown theon plots, data on acceleration, pressure level, and harvested energy appear linearly related. Calculated sound pressure level, and harvested energy appear topressure, be appear linearly Calculated correlation coefficients between acceleration and sound acceleration and harvested electric acceleration, sound pressure level, and harvested energy torelated. be linearly related.correlation Calculated coefficients between acceleration and sound pressure, acceleration and harvested electric energy, and sound pressure and harvested electric energy are overacceleration 0.976, as shown in Table 4.energy, This correlation coefficients between acceleration and sound pressure, and harvested electric and sound pressure and harvested electric energy are over 0.976, as shown in Table 4. This reveals that revealsand thatsound harvested energy, vibrations, and soundare pressure level as areshown strongly correlated energy, pressure and floor harvested electric energy over 0.976, in Table 4. This harvested energy, andfor sound pressure level can are strongly correlated with each other, with each other,floor andvibrations, a threshold harvested be employed detecting floor reveals that harvested energy, floor vibrations, andenergy sound pressure level areforstrongly correlated and a threshold for harvested energy can be employed for detecting floor vibrations and impact sound. vibrations and impact sound. with each other, and a threshold for harvested energy can be employed for detecting floor
vibrations and impact sound.
(a) Acceleration vs. SPL
(a) Acceleration vs. SPL
(b) Acceleration vs. Harvested Energy
(b) Acceleration vs. Harvested Energy
(c) SPL vs. Harvested Energy Figure10. 10.Correlation Correlationamong among slab slab acceleration, acceleration, SPL, Figure SPL,and andharvested harvestedenergy. energy.
(c) SPL vs. Harvested Energy Figure 10. Correlation among slab acceleration, SPL, and harvested energy.
Sensors 2018, 18, 4276 Sensors 2018, 18, x FOR PEER REVIEW
9 of 12 9 of 12
Table 4. Calculated correlation correlation coefficient. Table 4. Calculated coefficient.
Factors CorrelationCoefficient Coefficient Factors Correlation Acceleration vs. Sound Pressure Level 0.976 Acceleration vs. Sound Pressure Level 0.976 Acceleration vs. 0.993 Acceleration vs. Harvested HarvestedEnergy Energy 0.993 SoundPressure Pressure Level Level vs. 0.989 Sound vs.Harvested HarvestedEnergy Energy 0.989
4. Demonstration Demonstration of of the the Self-Powered Self-Powered Wireless Wireless Sensor 4. Sensor for for Floor Floor Vibration Vibration Detection Detection For feasibility feasibility demonstration, demonstration, aa self-powered self-powered wireless wireless sensor sensor and and mobile mobile application application software software are are For developed. Figure Figure 11 11 shows shows the the configuration configuration of of the the sensor. sensor. The The transmitter module is is composed of developed. transmitter module composed of one tuned tuned bimorph bimorph in in Figure Figure 5b 5b and The receiver receiver one and the the data data transmitting transmitting board, board, AmbioMote24(Tx) AmbioMote24(Tx) [23]. [23]. The module has has aa CPU and aa power Table 55 module CPU module, module, aa Bluetooth Bluetooth module, module, AmbioMote24(Rx), AmbioMote24(Rx), and power module. module. Table shows the the specifications specifications of of each each module. module.When Whenthe theenergy energyharvester harvestercharges chargeselectric electricenergy energy over shows over a a certain threshold due to abnormal floor vibrations, charged electric energy powers AmbioMote24(Tx) certain threshold due to abnormal floor vibrations, charged electric energy powers AmbioMote24(Tx) and sends sends aadatum datumtotoAmbioMote24(Rx). AmbioMote24(Rx). Next, CPU module activates the Bluetooth module, and Next, thethe CPU module activates the Bluetooth module, and andcommunicates it communicates a cellular phone, alarmsignal signalisis displayed displayed in in the the mobile it withwith a cellular phone, andand thethealarm mobile application software. software. application
Figure Figure 11. 11. Configuration Configuration of of the the self-powered self-powered wireless wireless sensor. sensor. Table 5. 5. Specifications Specifications of of the the self-powered self-powered wireless Table wireless sensor. sensor. Module Module
Main Specifications Main Specifications AmbioSYSTEMS AmbioMote24-A AmbioSYSTEMS AmbioMote24-A
AmbioMote24
AmbioMote24
Frequency band: 2.4 GHz Frequency band: 2.4 GHz ADC Convert resolution: 10 bit Capacitance: 0.5 µ ADC Convert resolution: 10 bit Communication distance: up to 80 m Capacitance: 0.5 μ 10 Hz Data transmission rate(Tx):
Communication distance: up to 80 m STMicroelectronics STM32L152RCT6 CPU Module
CPU Module Bluetooth Module
speed: max. 32 MHz10 Hz DataClock transmission rate(Tx):
Data bus width: 32 bit STMicroelectronics Memory: 256 KB, 32 KBSTM32L152RCT6 RAM, 8 KB ROM
Clock speed: max. 32+ MHz Firmtech FB155BC(SPP HID) Bluetooth Version: 2.1 (2.4 GHz ISM Band) Data bus width: 32 bit Communication distance: 10 m
Memory: 256 KB, 32 KB RAM, 8 KB ROM + HID) Figure 12 shows the test setup for the Firmtech feasibilityFB155BC(SPP demonstration. The transmitter module is Module 2.1 (2.42 GHz ISMfrom Band) installed on a bare Bluetooth slab, and the receiverBluetooth module isVersion: located about m away the center of the slab. A cellular phone is set by the receiver module. When impacting the slab10 with Communication distance: m a bang machine and an impact ball, the number of alarm signals displayed in the mobile application software is monitored. Figure 12 shows the test setup for the feasibility demonstration. The transmitter module is installed on a bare slab, and the receiver module is located about 2 m away from the center of the
Sensors 2018, 18, x FOR PEER REVIEW
10 of 12
slab. A cellular Sensors 2018, 18, 4276phone
is set by the receiver module. When impacting the slab with a bang machine 10 of 12 and an impact ball, the number of alarm signals displayed in the mobile application software is monitored. Table 6 summarizes the number of data transmissions caused by slab impacts. The Table 6 summarizes the number data transmissions slab impacts. harvested electric harvested electric energy from of a one-time impact ofcaused a bangby machine at the The center of the slab can energy from a one-time impact of a bang machine at the center of the slab can send seven alarm signals send seven alarm signals to a smartphone [25]. Here, one detection signal means one data to a smartphone [25].the Here, one detection signal means one data transmission fromenergy the transmitter transmission from transmitter module to the receiver module. The electric used for module to the receiver module. The electric energy used for one-time data transmission is about 11 µJis one-time data transmission is about 11 μJ (≈78.11 μJ/7 transmissions in case a 47 μF capacitor (≈ 78.11 µJ/7 transmissions in case a 47 µF capacitor is used). Differently, three successive impacts using used). Differently, three successive impacts using an impact ball can send four alarm signals [25]. an impact ball canthe send four alarm signals [25]. This is because the wireless transmitter, This is because wireless transmitter, AmbioMote24, initiates transmission whenAmbioMote24, the capacitor initiates transmission when the capacitor voltage is charged up to 5.3 V [10]. By using voltage is charged up to 5.3 V [10]. By using measured harvested energy in Table 3, wemeasured can check 2 5is
~ harvested Tablethe 3, we canmachine check thatatone with②~⑤ the bang at the that one energy impactinwith bang theimpact locations canmachine transmit onelocations signal. This can transmit one signal. This is because the bimorph can harvest about 11.71 µJ~12.38 µJ. Similarly, we because the bimorph can harvest about 11.71 μJ~12.38 μJ. Similarly, we estimate that 23 impacts 2 signals. 5 can transmit
~
estimate 23 impacts the impact locationsfour four Based on the with thethat impact ball at with locations ②~⑤ball canattransmit Based on thesignals. demonstration, it is demonstration, is confirmed that the cantilever-type piezoelectric harvester withcircuit the SEH confirmed that itthe cantilever-type piezoelectric harvester with the SEH interface caninterface generate circuit generate enough electric energy to send signals with one heavyweight impact of standard enoughcan electric energy to send wireless signals withwireless one impact of standard impact heavyweight impact sources. If the vibration level of an interested abnormal condition lower than sources. If the vibration level of an interested abnormal condition is lower than is the vibration the vibration induced by the standard heavyweight impact sources used, the harvestable will induced by the standard heavyweight impact sources used, the harvestable energy willenergy be smaller be smaller than in Table 3. For example, as described in [26], the impact forces of walking adults, than in Table 3. For example, as described in [26], the impact forces of walking adults, running running and jumping children are about 500 N, 1000 N, and N, respectively. shown children,children, and jumping children are about 500 N, 1000 N, and 30003000 N, respectively. AsAs shown in in Table1,1,the theimpact impactforce forceofofthe thebang bang machine machine is is about about 4000 4000 N. Table N. Based Based on on the the assumption assumption that thatthe the acceleration of of a slab is proportional to theto impact force and to the square root of harvested accelerationamplitude amplitude a slab is proportional the impact force and to the square root of 1(impact energy, the harvested energy of a walking location can belocation roughly estimated 1.22 µJ harvested energy, the harvested energyadult of a(impact walking adult ) ①) can beasroughly 2 2). Similarly, (78.11(µJ) the harvested energiesthe of running children andofjumping estimated×as(500(N)/4000(N)) 1.22 μJ (78.11(μJ)). ×Similarly, (500(N)/4000(N)) harvested energies running 1 children estimatedchildren as 4.88 µJ and µJ, respectively (impact location The applications childrencan andbejumping can be43.94 estimated as 4.88 μJ and 43.94 μJ, ). respectively (impact with higher thresholds than the harvested energies can be solved two system modifications: (1)with the location ①). The applications with higher thresholds than the with harvested energies can be solved amount of harvested energy can be easily multiplied by connecting a single energy harvester in parallel; two system modifications: (1) the amount of harvested energy can be easily multiplied by connecting and (2) the electric circuit design can be enhanced to minimize by considering the a single energy harvester in parallel; and (2) the electric circuit quiescent design canpower be enhanced to minimize sporadic nature of floor impacts. Wethe believe a study on the thresholds different applications quiescent power by considering sporadic nature of suitable floor impacts. Wefor believe a study on the should also be considered as important futureshould work. also be considered as important future work. suitable thresholds for different applications
Figure12. 12.Test Testsetup setupfor forfeasibility feasibilitydemonstration. demonstration. Figure Table 6. Impact condition and the number of data transmissions. Table 6. Impact condition and the number of data transmissions. Impact Source Source Impact Bang machine Bang machine Impact ball ball Impact
ImpactLocation Location Number Number Impacts Number Numberof ofSignals Signals Transmitted Impact ofofImpacts Transmitted 1
1 7 ① 1 7 1 ① 33 44
5.5.Conclusions Conclusions For Forthe therealization realizationof ofaasmart smartbuilding, building,Internet InternetofofThings Things(IoT) (IoT)applications applicationswith withconventional conventional acceleration been widely researched for implementing building automation with energy accelerationsensors sensorshave have been widely researched for implementing building automation with management technologies, and for resolving social issues, such as inter-floor noise-related disputes
Sensors 2018, 18, 4276
11 of 12
in apartments and the solitary death of an elderly person living alone. However, the installation of conventional accelerometers is sometimes controversial due to privacy invasion, especially if it is developed as a wireless system. To tackle this issue, this paper proposes a self-powered wireless sensor using a threshold-based method for the detection of floor vibrations. The vibration levels of the slab in the testbed are first measured by impacting the slab with the bang machine and the impact ball. By considering the vibration levels of the slab, the piezoelectric energy harvester is designed to generate electric energy over a certain threshold. A strong correlation among floor impact sound, slab vibrations, and electric energy harvested is revealed through the correlation study, and it is revealed that harvested electric energy can be used as a threshold to predict floor vibrations and the corresponding impact sound in buildings or apartments. Finally, the prototype of the self-powered wireless sensor is developed and its applicability is demonstrated. It is confirmed that the bimorph in Figure 5b with an SHE interface circuit can generate enough electric energy to send wireless signals by impacting with standard heavyweight impact sources. Due to the characteristics of the sensor—indirect measurement and less invasion of privacy—possible applications will be IoT systems to support a mediation process for inter-floor noise-related disputes in apartments with objective data on human-induced floor vibrations, or to detect the solitary death of an elderly person living alone by monitoring unusual patterns in floor vibrations. In addition, the benefits of using energy harvesters—easy installation and no maintenance costs to supply power in a transmitter—can expand the range of applications with technical improvements in designing vibration-based energy harvesting devices. Regarding future work, designing optimal electric circuits in consideration of floor vibration characteristics will be one of our research topics. For example, the development of electric circuits is required to minimize quiescent power by considering the sporadic nature of floor impacts [22], or to maximize harvestable power by using a decaying sinusoidal wave with a short duration, as shown in Figure 3. In addition, from a reliability point of view, various uncertainties in designing energy harvesters and in apartment environments, such as variation in the natural frequencies of slabs, should be considered [27]. Techniques for designing broadband energy harvesters may be a possible solution to these uncertainty problems. Finally, alarm thresholds and detection algorithms should be studied further. Author Contributions: B.C.J conceptualized this research. B.C.J and J.-W.P designed and performed the experiments. B.C.J and Y.C.H analyzed the experimental results. B.C.J wrote the draft manuscript. B.C.J. and Y.C.H edited and finalized the manuscript. Funding: This research was supported by the National Research Council of Science and Technology of Korea (Project Number: NK192F and NK213E). Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4. 5. 6. 7.
Kim, T.; Ramos, C.; Mohammed, S. Smart City and IoT. Future Gener. Comput. Syst. 2017, 76, 159–162. [CrossRef] Krishna, R.; Kummitha, R.; Crutzen, N. How do we understand smart cities? An evolutionary perspective. Cities 2017, 67, 43–52. Najafi-Ghalelou, A.; Zare, K.; Nojavan, S. Optimal scheduling of multi-smart buildings energy consumption considering power exchange capability. Sustain. Cities Soc. 2018, 41, 73–85. [CrossRef] Marinakis, V.; Doukas, H. An advanced IoT-based System for Intelligent Energy Management in Buildings. Sensors 2018, 18, 610. [CrossRef] [PubMed] Lam, M.; Mirshekari, M.; Pan, S.; Zhang, P.; Noh, H.Y. Robust Occupant Detection through Step-Induced Floor Vibration by Incorporating Structural Characteristics. Dyn. Coupled Struct. 2016, 4, 357–367. Mun, D.; Park, H.; Hwang, J. Prediction of Concrete Slab Acceleration and Floor Impact Noise Using Frequency Response Function. Trans. Korean Soc. Noise Vib. Eng. 2014, 24, 483–492. [CrossRef] Kim, K.Y.; Chung, J.Y.; Yeon, J.O.; Kim, M.J. Analysis of Floor Impact Sound Propagation around the Adjacent House. Appl. Mech. Mater. 2015, 752–753, 745–748. [CrossRef]
Sensors 2018, 18, 4276
8. 9. 10. 11. 12. 13.
14. 15.
16. 17. 18. 19. 20.
21.
22.
23. 24. 25. 26. 27.
12 of 12
Lee, P.; Jeong, J.; Park, J.; Jeon, J. Comparison of Standard Floor Impact Sources with Human Impact Source. Trans. Korean Soc. Noise Vib. Eng. 2006, 8, 789–796. Jeon, J.Y.; Ryu, J.K.; Jeong, J.H.; Tachibana, H. Review of the Impact Ball in Evaluating Floor Impact Sound. Acta Acust. United Acust. 2006, 92, 777–786. Lee, S.; Youn, B.D.; Jung, B.D. Robust segment-type energy harvester and its application to a wireless sensor. Smart Mater. Struct. 2009, 18, 095021. [CrossRef] Roundy, S.; Wright, P.K.; Rabaey, J.M. Energy Scavenging for Wireless Sensor Networks: With Special Focus on Vibrations; Springer: Berlin, Germany, 2004. Kim, H.S.; Kim, J.-H.; Kim, J. A review of piezoelectric energy harvesting based on vibration. Int. J. Precis. Eng. Manuf. 2011, 12, 1129–1141. [CrossRef] Yoon, H.; Youn, B.D.; Kim, H.S. Kirchhoff plate theory based electromechanically-coupled analytical model considering inertia and stiffness effects of a surface-bonded piezoelectric patch. Smart Mater. Struct. 2016, 25, 025017. [CrossRef] Erturk, A.; Inman, D.J. An experimentally validated bimorph cantilever model for piezoelectric energy harvesting from base excitations. Smart Mater. Struct. 2009, 18, 25009–25018. [CrossRef] Kim, J.E.; Kim, H.; Yoon, H.; Kim, Y.Y.; Youn, B.D. An Energy conversion model for cantilevered piezoelectric vibration energy harvesters using only measurable parameters. Int. J. Precis. Eng. Manuf.-Green Technol. 2015, 2, 51–57. [CrossRef] Roundy, S.; Kright, P.K. A piezoelectric vibration based generator for wireless electronics. Smart Mater. Struct. 2004, 13, 1131–1142. [CrossRef] Sodano, H.A.; Park, G.; Inman, D.J. Estimation of Electric Charge Output for Piezoelectric Energy Harvesting. Strain 2004, 40, 49–58. [CrossRef] Chen, S.N.; Wang, G.J.; Chien, M.C. Analytical modeling of piezoelectric vibration-induced micro power generator. Mechatronics 2006, 16, 379–387. [CrossRef] Goldschmidtboeing, F.; Woias, P. Characterization of different beam shapes for piezoelectric energy harvesting. J. Micromech. Microeng. 2008, 18, 104013. [CrossRef] Hehn, T.; Hagedorn, F.; Maurath, D.; Marinkovic, D.; Kuehne, I.; Frey, A.; Manoli, Y. A Fully Autonomous Integrated Interface Circuit for Piezoelectric Harvesters. IEEE J. Solid-State Circuits 2012, 47, 2185–2198. [CrossRef] Gasnier, P.; Willemin, J.; Boisseau, S.; Despesse, G.; Condemine, C.; Gouvernet, G.; Chaillout, J. An Autonomous Piezoelectric Energy Harvesting IC Based on a Synchronous Multi-Shot Technique. IEEE J. Solid-State Circuits 2014, 49, 1561–1570. [CrossRef] Quelen, A.; Morel, A.; Gasnier, P.; Grézaud, R.; Monfray, S.; Pillonnet, G. A 30 nA quiescent 80 nW to 14 mW power-range shock-optimized SECE-based piezoelectric harvesting interface with 420% harvested energy improvement. In Proceedings of the IEEE International Solid State Circuits Conference, San Francisco, CA, USA, 11–15 February 2018; pp. 150–152. Analog Sensor Node. Available online: http://www.ambiosystems.com/ (accessed on 22 July 2018). Piezo Systems Product Catalog. Available online: http://piezo.com/ (accessed on 22 July 2018). Video of Feasibility Demonstration. Available online: https://youtu.be/fliSTD6A0G8 (accessed on 31 August 2018). Jeon, J.Y.; Lee, P.J.; Sato, S. Use of the standard rubber ball as an impact source with heavyweight concrete floors. J. Acoust. Soc. Am. 2009, 126, 167–178. [CrossRef] [PubMed] Jung, B.C.; Yoon, H.; Oh, H.; Lee, G.; Yoo, M.; Youn, B.D.; Huh, Y.C. Hierarchical model calibration for designing piezoelectric energy harvester in the presence of variability in material properties and geometry. Struct. Multidiscip. Optim. 2016, 53, 161–173. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
