sensors Article
Two Capacitive Micro-Machined Ultrasonic Transducers for Wind Speed Measurement Gia Thinh Bui, Yu-Tsung Jiang and Da-Chen Pang * Department of Mechanical Engineering, National Kaohsiung University of Applied Sciences, 415 Jiangong Road, Sanmin District, Kaohsiung 807, Taiwan; [email protected] (G.T.B.); [email protected] (Y.J.) * Correspondence: [email protected]; Tel.: +886-7-381-4526 (ext. 5331) Academic Editor: Xiaoning Jiang Received: 7 April 2016; Accepted: 30 May 2016; Published: 2 June 2016
Abstract: This paper presents a new wind speed measurement method using a single capacitive micro-machined ultrasonic transducer (CMUT). The CMUT was arranged perpendicular to the direction of the wind flow, and a reflector was set up a short distance away, facing the CMUT. To reduce the size, weight, cost, and power consumption of conventional ultrasonic anemometers this study proposes two CMUT designs for the measurement of wind speed using either the amplitude of the signal or the time of flight (TOF). Each CMUT with a double array element design can transmit and receive signals in five different operation modes. Experiments showed that the two CMUT designs utilizing the TOF were better than those utilizing the amplitude of the signal for wind speed measurements ranging from 1 m/s to 10 m/s, providing a measurement error of less than 0.2 m/s. These results indicate that the sensitivity of the TOF is independent of the five operation modes. Keywords: capacitive micro-machined ultrasonic transducer (CMUT); wind speed measurement; amplitude; time-of-flight
1. Introduction Capacitive micro-machined ultrasonic transducers (CMUTs) emerged as an interesting alternative to piezoelectric transducers in the mid-1990s [1,2]. Over the past two decades, several research studies have been conducted to explore the fabrication [3,4], characterization [5], and modeling [6] of CMUTs. Our goal in conducting this research was to develop new CMUT array designs that are useful for measuring wind speed through the use of a single CMUT. The measurement of fluid speed, particularly wind speed, using ultrasonic transducers has been found to have important scientific and industrial applications [7,8]. Wind speed measurements can be applied not only in industrial fields, but also for the purpose of environmental monitoring and the control of various systems, such as indoor air conditioning systems, weather forecasting systems, and other sensors [9,10]. Some applications, such as those providing an evaluative measurement of wind speed for wind power generation, require measurements with low levels of uncertainty [11]. To design an instrument for measuring wind speed with measurement uncertainty that is as low as possible, a detailed study of the sources of measurement uncertainties and their propagation was conducted [12]. A common approach for determining wind speed is to measure the time-of-flight (TOF) by means of an ultrasonic signal, where the TOF is defined as the time required for an ultrasonic wave to travel from its source of transmission to the receiving transducer [13]. In the existing literature, several measurement techniques based on TOF estimates have been presented and discussed, including the threshold detection (TH), phase difference (PD), cross-correlation, maximum likelihood estimation (MLE), Kalman filter and wavelet transform techniques [14]. The TH and PD techniques are straightforward when compared with other techniques that make use of signal correlation [15]. However, an estimation of the TOF requires the consideration of factors such as attenuation of Sensors 2016, 16, 814; doi:10.3390/s16060814
www.mdpi.com/journal/sensors
Sensors 2016, 16, 814
2 of 9
the medium, random noise, and reflections, which lower the signal noise ratio (SNR) and increase uncertainty to high levels. To develop a minimally complex CMUT array for ultrasound transmissions and reception, some publications recommended an ultrasound flow measurement system with a transmitter and receiver on the same substrate as our system. Instead of using Doppler effects for flow measurement [16,17], we used the TOF method to determine wind velocity. In the comparison with the methods used in previous studies [18,19], our method also has the advantage of allowing for the quick calculation of results from simple equations. In this paper, we propose two CMUT designs to measure wind speed, based on the TOF and amplitude of ultrasonic signals between transmission and reception. Specifically, we use only a single CMUT to yield a measurement of velocity with small spatial resolution of 10 mm between the reflector and transducer. The findings from this study can reduce the number of sensors needed, compared with previous systems [20–22]. Furthermore, the 3 mm ˆ 3 mm CMUT presented in this study is also smaller than other ultrasonic anemometers that have been used previously [23]. Because of the smaller distance between the CMUT and reflector, the travel time is shorter and the CMUT’s sampling rate is higher. The CMUTs for wind speed measurement have a novel design for the transmission and reception of signals in five different modes of operation. This approach also uses the available array element area efficiently. In this paper, the implementation of two new CMUT designs for wind speed testing is described, and the pulse-echo testing results are discussed. These results show that a single CMUT can be used for wind speed measurement with good accuracy. In addition, the whole system can be realized with a simple structure and low costs. 2. CMUT Design The basic mechanical structure of a CMUT is similar to that of a capacitor with two parallel plates (see Figure 1). The membrane, filling layer, and side wall are all made of polymer, an epoxy-type photoresist (SU-8). The filling layer is designed to protect the gold top electrode from the ambience and external corrosion during regular operation. A thin indium tin oxide bottom electrode is adhered on a substrate and separated from the membrane by an air cavity. The geometric parameters of a CMUT are summarized Sensors 2016, 16, 814 in Table 1. 3 of 10
Figure 1. Cross-sectional schematic of a single CMUT cell. Figure 1. Cross-sectional schematic of a single CMUT cell. Table 1. Mechanical dimensions of a CMUT cell. Table 1. Mechanical dimensions of a CMUT cell. Parameter Measurement (μm) Membrane diameter 140 Parameter Measurement (µm) Membrane thickness 5 Membrane diameter 140 Sidewall width 105 Membrane thickness Cavitywidth height 210 Sidewall Bottom electrode 0.05 Cavity heightthickness 2 Bottom electrode thickness 0.05 Top electrode diameter 108 Top diameter 108 Topelectrode electrode thickness 0.15 Top electrode thickness
0.15
Sensors 2016, 16, 814
3 of 9
Figure 1. Cross-sectional schematic of a single CMUT cell.
The aim of this research is to design new CMUTs with of divergent characteristics. The transmission Figure 1. Cross-sectional schematic a single CMUT cell. and reception modes of a single device can be operated separately so that the element can be adjusted Table 1. Mechanical dimensions of a CMUT cell. 1. ultrasonic Mechanicalsignals. dimensions a CMUT cell.as designed can work with individually to transmit andTable receive The of single device Parameter Measurement (μm) five operation modes using different ratio of transmission and reception cells. In this study, two Parameter Measurement (μm) Membrane 140 CMUT designs are presented, as showndiameter in Figure 2. Each one was arranged in a square area measuring Membrane diameter 140 Membrane thicknessof 140 and 5 µm, respectively, 5 3 mm ˆ 3 mm with a diameter and thickness for each cell membrane Membrane thickness 5 Sidewall width 10 and a cavity height of 2 µm. Each CMUT has two arrays of cells, called the majority and minority Sidewall width 10 Cavity height 2 arrays. The majority array has more cells than the minority array, with2a 2:1 ratio, in these designs. In Cavity height Bottom electrode thickness 0.05 the wave-type CMUT design, the array element is separated using 0.1 mm spacing to include 416 cells Bottom electrode thickness 0.05 Top electrode diameter 108 inside it, as shown in Figure Top 2a. Figure 2b shows the parallel-line type CMUT arranged in a rectangular electrode diameter 108 Top electrode thickness 0.15 array, with a majority arrayTop of 80 cells arranged electrode thicknessin a 3 mm ˆ 0.65 mm 0.15area and a minority array of 40 cells arranged in a 3 mm ˆ 0.35 mm area.
(a) (a)
(b) (b)
Figure 2. Diagram of two CMUT designs: (a) Wave-type; (b) Parallel-line type. Figure of two twoCMUT CMUTdesigns: designs:(a) (a)Wave-type; Wave-type; Parallel-line type. Figure 2. 2. Diagram Diagram of (b)(b) Parallel-line type.
3. Measurement Principle 3.3.Measurement MeasurementPrinciple Principle It is common to use two ultrasonic transducers for measuring wind speed by lining up the ItItisis common to use two two ultrasonic ultrasonictransducers transducersfor formeasuring measuring wind speed lining common to use speed by by lining up up the the transmitting and receiving transducers at a specific angle. This study wind proposes a new measurement transmitting and receiving transducers at a specific angle. This study proposes a new measurement transmitting and receiving transducers at a specific angle. This study proposes a new measurement system that uses only one CMUT and one reflector within a small space as shown in Figure 3. In this system CMUT and andone onereflector reflectorwithin within a small space shown in Figure In this systemthat thatuses uses only only one CMUT a small space as as shown in Figure 3. In3.this design, the wind flows perpendicular to the ultrasound signal path between the CMUT and the design, perpendicular to tothe theultrasound ultrasoundsignal signalpath path between CMUT design,the the wind wind flows flows perpendicular between thethe CMUT andand the the reflector. The wind will change the ultrasound propagation by increasing the TOF and reducing the reflector.The The wind wind will change increasing thethe TOF andand reducing the the reflector. change the theultrasound ultrasoundpropagation propagationbyby increasing TOF reducing amplitude of the ultrasound signals. As shown in Figure 3, the red vector is the ultrasonic path when amplitudeof ofthe the ultrasound ultrasound signals. red vector is the ultrasonic path when amplitude signals.As Asshown shownininFigure Figure3,3,the the red vector is the ultrasonic path when it is altered by the wind flow, d is the distance (i.e., 10 mm) between the CMUT and the reflector, and alteredby bythe the wind wind flow, flow, dd is thethe CMUT and thethe reflector, andand ititisisaltered isthe thedistance distance(i.e., (i.e.,1010mm) mm)between between CMUT and reflector, θ isθ the angle between the transmitted path of the ultrasonic signal and andthe theshifted shiftedcondition conditionofof the is the angle between the transmitted path of the ultrasonic signal θ is the angle between the transmitted path of the ultrasonic signal and the shifted conditionthe of the wind direction. winddirection. direction. wind
Figure the CMUT. CMUT. Figure3.3.Wind Windspeed speedmeasurement measurement using the Figure 3. Wind speed measurement using the CMUT.
3.1. Measuring Wind Velocity In this study, a method of measuring wind velocity was developed based on the different measurements times of the transmission time of the ultrasonic signals. This measured time difference ∆t is expressed as Equation (6), which is derived from Equations (4) and (5). The latter equations are
measurements times of the transmission time of the ultrasonic signals. This measured time difference Δt is expressed as Equation (6), which is derived from Equations (4) and (5). The latter equations are themselves derived from Equations (2) and (3) in terms of the velocity of wind v being much lower than sound speed c. The time difference of the ultrasonic signals transmitted Δt1 and echoed Δt2 were determined by substituting the measured transmitted time t1 and echo time t2. The Sensors 2016, 16, 814 4 of 9 transmission time of the ultrasonic signal t0 to the reflector without wind is measured as presented in Equation (1): themselves derived from Equations (2) and (3) in terms of the velocity of wind v being much lower d t than sound speed c. The time difference of the ultrasonic signals transmitted ∆t1 and echoed ∆t2 were(1) c determined by substituting the measured transmitted time t1 and echo time t2 . The transmission time of the ultrasonic signal t0 to the reflector without wind is measured as presented in Equation (1):
t
d
vt
t
c
1
v c
1 v d 2 c t0 “ t
t
1 v 2 c
t
⋯
1 v 2 c
t
c
tt1 “ b
b dd2 ` pvt 2vt0 q2 cc
d2 ` p2vt0 q2
c ´ v2v ¯2 ´1 ¯ ´1 ¯2v ´1 ¯2v 2v 2 4 n t t` 1 v t0 ` t 1 v t0 `t . . . ` ⋯1 v t0 t “ t0 11` “ t 0 cc 2 2c c 2 2c c 2 2c c
(3)
(2)
d
˙ ˆ ˙ ˆ ˙ ˆ ˙ 1 2v1 v2 1 2v 4 1 2v n 2v 2 t0 ` t0 ` . . . ` t0 t ∆t “ t0 ` t c c 2 c2 c 2 c 2 c 1 ´ v ¯2 ∆t1 “ t1 ´ t0 « 1 2v t0 t t 2 c t ∆t ˆ2 c˙2 1 2v ∆t2 “ t2 ´ t0 « t0 2 5c v ∆t ∆t ∆t c2 d 5 ´2v ¯ ∆t “ ∆t1 ` ∆t2 “ Temperature and humidity affect the density of the 2 cair, ccausing changes in the velocity t2 “
(2) (1)
ˆ
“ t0
1`
(3)(4) (4)
(5)
(5)
(6) (6)
of the ultrasound signal [24–26]. If the CMUT used at room temperature less than 1 °C variation, Temperature and humidity affectisthe density of the air, causingwith changes in the velocity of thethe ˝ ultrasound velocity change is within 0.2%. The humidity effect on ultrasound velocity is the even ultrasound signal [24–26]. If the CMUT is used at room temperature with less than 1 C variation, smaller. The ultrasound velocity change only 0.05% ifeffect the temperature constant and humidity ultrasound velocity change is within 0.2%.isThe humidity on ultrasoundisvelocity is even smaller. The ultrasound velocity change is only 0.05% if the temperature is constant and humidity changes changes remain within 10%. remain within 10%.
3.2. Wind Speed Measurement System Configuration 3.2. Wind Speed Measurement System Configuration
A wind speed measurement system was initiated for basic measurements, as shown in Figure 4. A wind speed measurement system was initiated for basic measurements, as shown in Figure 4. The measurement box had the dimensions of 120 cm × 60 cm × 60 cm. On cross-sections The measurement box had the dimensions of 120 cm ˆ 60 cm ˆ 60 cm. On cross-sections perpendicular perpendicular to the wind flow direction, the wind velocity was determined on a grid. A transducer to the wind flow direction, the wind velocity was determined on a grid. A transducer was installed on was installed on the calibration system. The wind speed tests were performed in the laboratory at a the calibration system. The wind speed tests were performed in the laboratory at a temperature of temperature of 25 °C ± 0.5 °C and humidity of 60% ± 3%. The effects of temperature and humidity on 25 ˝ C ˘ 0.5 ˝ C and humidity of 60% ˘ 3%. The effects of temperature and humidity on ultrasound ultrasound velocity were very small. The environment uncertainty was approximately 0.2%. velocity were very small. The environment uncertainty was approximately 0.2%.
Figure4.4.Photograph Photographof ofthe thewind wind speed measurement Figure measurementconfiguration. configuration.
Figure 5 shows the setup for the wind speed experiment. The CMUT was powered by DC 100 V and AC 300 V power sources, which produced the ultrasonic pulse signals used in the experiment. A generator (DPR 300 Pulse/Receiver, JSR Ultrasonics, Pittsford, NY, USA) transmits a short pulse, which has an amplitude of 0–475 V. A variable DC bias voltage of 0–300 V (N5751A DC Power Supply, Keysight Technologies, Santa Rosa, CA, USA) is superimposed on the transient voltage. The receiver signals were obtained with an oscilloscope (DPO 4104B Digital Phosphor Oscilloscope, Tektronix,
Figure 5 shows the setup for the wind speed experiment. The CMUT was powered by DC 100 V Sensors 2016, 16, 814 5 of 10 and AC 300 V power sources, which produced the ultrasonic pulse signals used in the experiment. A generator (DPR 300 Pulse/Receiver, Ultrasonics, Pittsford, USA) a DC short Figure 5 shows the setup for theJSR wind speed experiment. The NY, CMUT was transmits powered by 100pulse, V which an V amplitude of 0–475 V.produced A variable bias voltage of 0–300 DC Power and has AC 300 power sources, which the DC ultrasonic pulse signals usedV in (N5751A the experiment. A Sensors 2016, 16, 814 5 of 9 generator (DPRTechnologies, 300 Pulse/Receiver, Ultrasonics, NY, USA) a short pulse,The Supply, Keysight Santa JSR Rosa, CA, USA)Pittsford, is superimposed ontransmits the transient voltage. whichsignals has an were amplitude of 0–475 A variable DC bias voltage 0–300 VPhosphor (N5751A DC Power receiver obtained withV.an oscilloscope (DPO 4104Bof Digital Oscilloscope, Supply, Keysight Technologies, Santa Rosa, CA, USA) is superimposed on the transient voltage. The Tektronix, OR, Beaverton, OR, USA)toconnected the TOFs CMUT. Any TOFs in the measurements Beaverton, USA) connected the CMUT.toAny obtained in theobtained measurements were recorded receiver signals were obtained with an oscilloscope (DPO 4104B Digital Phosphor Oscilloscope, weretransmitted recorded and and to atransmitted computer. to a computer. Tektronix, Beaverton, OR, USA) connected to the CMUT. Any TOFs obtained in the measurements were recorded and transmitted to a computer. Receiver output T/R
Receiver output
TRIG/SYNC
Through
T/R
Pulse/Receiver Through Generator
Oscilloscope
TRIG/SYNC
Pulse/Receiver Generator
Bias tee
DC Power Supply
Bias tee
DC Power Supply
Oscilloscope
Ch1
CMUT (Transmitter) CMUT (Transmitter)
CMUT (Receiver)
Ch1
EXT/TRIG
EXT/TRIG
Bias tee
CMUT Bias (Receiver) tee Ultrasound
Reflector
DC Power Supply DC Power Supply
Ultrasound
Reflector
5. Setup for wind speed experiment. Figure 5. Figure 5. Setup for wind speed experiment.
Figure 66 shows shows the the received received ultrasonic signal’s signal’s amplitude. The The measurement measurement system system starts starts its its Figure Figure 6 shows the receivedultrasonic ultrasonic signal’s amplitude. amplitude. The measurement system starts its operation with a pulse-wave transmitted and received by the transducer, which is propagated in the operation with a pulse-wave bythe thetransducer, transducer,which which is propagated in the operation with a pulse-wavetransmitted transmittedand and received received by is propagated in the medium. The time needed for the detection of the received ultrasonic signal was determined using a medium. The time needed ofthe thereceived receivedultrasonic ultrasonic signal determined medium. The time neededfor forthe thedetection detection of signal waswas determined usingusing a reference threshold. From this detection, a data acquisition process utilizing the transmitted and a reference dataacquisition acquisitionprocess processutilizing utilizing transmitted referencethreshold. threshold. From From this detection, detection, aa data thethe transmitted andand received signals was conducted to determine any and all detection times. Then, an analysis of the the received signals was conductedtotodetermine determine any any and all anan analysis of the received signals was conducted all detection detectiontimes. times.Then, Then, analysis of signals ledled to estimates estimates ofof the PD between the transmitted andreceived receivedsignals. signals. The number of cells cells signals to estimates thePD PDbetween betweenthe the transmitted transmitted and The number of cells signals led to of the and received signals. The number of and the signal amplitudes of the two CMUT designs in five operation modes are shown in Table signal amplitudes thetwo twoCMUT CMUTdesigns designs in in five shown in in Table 2. 2.2. andand the the signal amplitudes ofofthe five operation operationmodes modesare are shown Table
Figure Timeresponse responseof ofpulse-echo pulse-echo signal mode. Figure 6. 6. Time signalin infull/full full/full mode.
Figure 6. Time response of pulse-echo signal in full/full mode. The operation of the transducer is based on the ratio of transmission and reception cells. The Table 2. Number of cells and signal amplitudes of the two CMUT designs in the five operation modes. experimental measurements have five modes: full transmission/full reception, majority The operation of the transducer is based on the ratio of transmission and reception cells. The transmission/ majority reception, majority reception, minority Wave-Typetransmission/minority Design Parallel-Line Type Design
experimental measurements have five modes: full transmission/full reception, majority Mode Signal Transmission/Reception Signal transmission/ majority reception,Transmission/Reception majority transmission/minority reception, minority (Cell Number) (mV) (Cell Number) (mV) Full transmission/Full reception Majority transmission/Majority reception Majority transmission/Minority reception Minority transmission/Majority reception Minority transmission/Minority reception
416/416 274/274 274/142 142/274 142/142
634 362 226 244 158
360/360 240/240 240/120 120/240 120/120
546 324 196 214 140
transmission/full reception mode had the maximum received signals and the highest sensitivity. The minority transmission/majority reception mode, had fewer transmitted signals than the majority transmission/majority reception mode but still received some signals and was considered better and relatively more efficient. Sensors 2016, 16, 814
Table 2. Number of cells and signal amplitudes of the two CMUT designs in the five operation modes.
6 of 9
The operation of the transducer is basedWave-Type on the ratio of transmission and reception cells. The Design Parallel-Line Type Design experimental measurements have five modes: full transmission/full reception, majority transmission/ Mode Transmission/Reception Signal Transmission/Reception Signal (Cell Number) reception, (mV)minority (Cell Number) (mV) majority reception, majority transmission/minority transmission/majority Full transmission/Full reception 416/416 634 360/360 546 reception, and minority transmission/minority reception. The full transmission/full reception mode Majority transmission/Majority reception 274/274 362 240/240 324 had the maximum received signals and the highest sensitivity. The minority transmission/majority Majority transmission/Minority reception 274/142 226 240/120 196 reception mode, had fewerreception transmitted signals than the majority reception Minority transmission/Majority 142/274 244 transmission/majority 120/240 214 Minority 142/142 158 relatively 120/120 mode buttransmission/Minority still received somereception signals and was considered better and more efficient. 140 4. ExperimentalResults Results 4. Experimental ˝ In this experiment the transducer has a 90 90° directivity angle and a reflecting surface set at a distance of 10 mm. The maximum signal of the first echo was used as an observation point, whereby it was found that the attenuation of the signal increases as the wind wind speed speed increases. increases. The peak time increases slowly along with the increasing wind speed, but the operation time of the first echo and of the resonance took advantage of the time difference between the resonancefrequency frequencydoes doesnot notchange. change.This Thisstudy study took advantage of the time difference between increasing attenuation of the of echo determine wind speed changes. shown inAs Figure 7, the the increasing attenuation thesignals echo to signals to determine wind speedAschanges. shown in experiments measured wind speeds from 1 m/s to 10 m/s and the results of the full transmission/full Figure 7, the experiments measured wind speeds from 1 m/s to 10 m/s and the results of the full reception mode and minoritymode transmission/majority reception showed thatreception the valuesshowed of the maximum transmission/full reception and minority transmission/majority that the points signal points changeofwhen the wind increases. values of of the the echo maximum the echo signalspeed change when the wind speed increases.
(a)
(b)
Figure 7. 7. Increasing to 10 m/s: (a) The full Figure Increasing attenuation attenuationof ofthe theecho echosignal signalby bywind windspeeds speedsfrom from1 m/s 1 m/s to 10 m/s: (a) The transmission/full reception modemode results; (b) The transmission/majority reception mode full transmission/full reception results; (b) minority The minority transmission/majority reception results. mode results.
4.1. Using Amplitude of Signal to Measure Wind Speed 4.1. Using Amplitude of Signal to Measure Wind Speed As shown in Figure 8a, the wave-type transducer design was tested. The experimental results of As shown in Figure 8a, the wave-type transducer design was tested. The experimental results the five modes were affected by the wind speed, with sensitivity of magnitude signals and a of the five modes were affected by the wind speed, with sensitivity of magnitude signals and a standard deviation of error as shown in Table 3. Figure 8b shows the results achieved when using standard deviation of error as shown in Table 3. Figure 8b shows the results achieved when using the parallel-line type ultrasonic transducer design for measurement. The measurement results of the the parallel-line type ultrasonic transducer design for measurement. The measurement results of the five modes were affected by the wind speed, with attenuation sensitivity of magnitude signals and a five modes were affected by the wind speed, with attenuation sensitivity of magnitude signals and a standard deviation of error as shown in Table 3. The results were based on the least square method standard deviation of error as shown in Table 3. The results were based on the least square method to to draw the best fit line. From the results, it can be concluded that the strength of the wind speed had draw the best fit line. From the results, it can be concluded that the strength of the wind speed had a a direct effect on the amplitude of the ultrasonic signals. The ultrasonic echo signals of the two types direct effect on the amplitude of the ultrasonic signals. The ultrasonic echo signals of the two types of of transducers were smaller when the attenuation of the wind speed increased. The measurements of transducers were smaller when the attenuation of the wind speed increased. The measurements of the two transducers in full transmission/full reception mode demonstrated high sensitivity significance. It was found that the minority transmission/majority reception mode had a larger signal than in majority transmission/minority reception mode.
Sensors 2016, 16, 814
7 of 10
the two transducers in full transmission/full reception mode demonstrated high sensitivity significance. It was found that the minority transmission/majority reception mode had a larger Sensors 2016, 16, 814 7 of 9 signal than in majority transmission/minority reception mode.
(a)
(b)
Figure 8. Effect of wind speed on measurement results: (a) Wave-type CMUT design; (b) Parallel-line Figure 8. Effect of wind speed on measurement results: (a) Wave-type CMUT design; (b) Parallel-line type CMUT design. type CMUT design. Table of measurement measurement for for the the two two CMUT CMUT designs. designs. Table 3. 3. The The sensitivities sensitivities and and standard standard deviation deviation errors errors of Designs Designs
Wave-type CMUT Wave-type CMUT
Parallel-line Parallel-line type type CMUT CMUT
Modes Modes Full/Full Full/Full Majority/Majority Majority/Majority Majority/Minority Majority/Minority Minority/Majority Minority/Majority Minority/Minority Minority/Minority Full/Full Full/Full Majority/Majority Majority/Majority Majority/Minority Majority/Minority Minority/Majority Minority/Majority Minority/Minority Minority/Minority
Amplitude TOF Amplitude TOF Sensitivity Sensitivity Error (m/s) Sensitivity−1 2 Error (m/s) Sensitivity Error Error (mV/ms−1) (ns/(ms ) ) (mV/ms´1 ) (m/s) (ns/(ms´1 )2 ) (m/s) 8.807 0.198 0.617 0.153 8.807 0.198 0.617 0.153 8.379 0.254 0.616 0.154 8.379 0.254 0.616 0.154 7.161 0.235 0.614 0.168 7.161 0.235 0.614 0.168 7.555 0.272 0.617 0.161 7.555 0.272 0.617 0.161 6.846 0.330 0.616 0.165 6.846 0.330 0.616 0.165 9.906 0.189 0.618 0.155 9.906 0.189 0.618 0.155 9.552 0.180 0.616 0.156 9.552 0.180 0.616 0.156 8.944 0.249 0.615 0.170 8.944 0.249 0.615 0.170 9.166 0.241 0.617 0.165 9.166 0.241 0.617 0.165 8.171 0.288 0.617 0.172 8.171 0.288 0.617 0.172
4.2. Using Differences in TOF’s to Measure Wind Speed 4.2. Using Differences in TOF’s to Measure Wind Speed In this study, the transducers used self-transmitted and self-received signals sent over a reflector In this study, the transducers used self-transmitted and self-received signals sent over a distance of 10 mm to measure wind speed. The influence of the wind causes the point of a signal’s reflector distance of 10 mm to measure wind speed. The influence of the wind causes the point of a arrival on the reflector to be offset by a distance of vt, and the wind effect on the speed of an ultrasound signal’s arrival on the reflector to be offset by a distance of vt, and the wind effect on the speed of an signal is offset again after reflection as shown in Figure 3. The transducers were used in experiments ultrasound signal is offset again after reflection as shown in Figure 3. The transducers were used in to measure wind speeds ranging from 1 m/s to 10 m/s. With greater wind speeds, the time required experiments to measure wind speeds ranging from 1 m/s to 10 m/s. With greater wind speeds, the to receive signals increases. The TOF can be calculated from the experimental results and compared time required to receive signals increases. The TOF can be calculated from the experimental results with the 2016, theoretical 16, 814 results as shown in Figure 9. 8 of 10 andSensors compared with the theoretical results as shown in Figure 9. There are significant errors evident when the wind speed is small. This contradicts the presented theoretical values. The results utilized the least square method to draw the second order curve fitting. Table 3 shows the wind speed measurement error using a difference in TOF that is slightly lower than the amplitude of the signal.
(a)
(b)
Figure 9. Comparison of the results from the theoretical calculation and the experimental Figure 9. Comparison of the results from the theoretical calculation and the experimental measurements measurements according to different TOFs: (a) Wave-type CMUT design; (b) Parallel-line type according to different TOFs: (a) Wave-type CMUT design; (b) Parallel-line type CMUT design. CMUT design.
5. Conclusions In this paper, two designs of a CMUT with five modes that can be used to measure the velocity of wind blowing perpendicular to the CMUT were presented. In addition, analyses of the ultrasound signal amplitudes and the TOF differences due to different wind speeds were conducted. The
Sensors 2016, 16, 814
8 of 9
There are significant errors evident when the wind speed is small. This contradicts the presented theoretical values. The results utilized the least square method to draw the second order curve fitting. Table 3 shows the wind speed measurement error using a difference in TOF that is slightly lower than the amplitude of the signal. 5. Conclusions In this paper, two designs of a CMUT with five modes that can be used to measure the velocity of wind blowing perpendicular to the CMUT were presented. In addition, analyses of the ultrasound signal amplitudes and the TOF differences due to different wind speeds were conducted. The problems with the measurement range for the TOF differences and signal amplitudes of the continuous-wave CMUT were compared. The two CMUT designs that used the TOF method did not show much of a difference in terms of sensitivity and error. The TOF determination method can be used to measure wind speeds without calibration. Also, it was observed that the full transmission/full reception mode was the best choice in terms of providing greater sensitivity and less uncertainty in wind speed measurement when using the amplitude of signal method. The minority transmission/majority reception mode is proposed for good energy efficiency since the design can be selected for signal transmission. The parallel-line type design produced results that were only slightly better than those of the wave-type design. Thus, the two CMUT designs can be used to build a wind speed measurement system that is independent of the five modes and has the advantages of high accuracy, low cost, and easy portability, as well as great potential for future applications. Acknowledgments: The authors would like to acknowledge the financial support of MEMS and Precision Machinery Research and Development Center at National Kaohsiung University of Applied Sciences for the development of the CMUTs. We would also like to acknowledge the Ministry of Sciences and Technology (MOST) for its partial support under project number MOST 104-2221-E-151-012. The wind speed experiment was performed at the Thermo-Fluid-Optical Visualization laboratory of the Department of Mold and Die Engineering at National Kaohsiung University of Applied Sciences. Author Contributions: Da Chen Pang and Yu Tsung Jiang conceived and designed the experiments; Yu Tsung Jiang performed the experiments; Da Chen Pang and Gia Thinh Bui analyzed the data; Gia Thinh Bui and Da Chen Pang wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.
References 1. 2.
3.
4. 5.
6.
Haller, M.I.; Khuri-Yakub, B.T. A surface micro-machined electrostatic ultrasonic air transducer. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 1996, 43, 1–6. [CrossRef] Ladabaum, I.; Khuri-Yakub, B.T.; Spoliansky, D.; Haller, M.I. Micro-machined ultrasonic transducers (MUTs). In Proceedings of the 1995 IEEE Ultrasonics Symposium, Seattle, WA, USA, 7–10 November 1995; Volume 1, pp. 501–504. Jin, X.C.; Ladabaum, I.; Degertekin, F.L.; Calmes, S.; KhuriYakub, B.T. Fabrication and characterization of surface surface micromachined capacitive ultrasonic immersion transducers. J. Microelectromech. Syst. 1999, 8, 100–114. Huang, Y.; Ergun, A.S.; Haeggstrom, E.; Badi, M.H.; KhuriYakub, B.T. Fabricating capacitive micro-machined ultrasonic transducers with wafer-bonding. J. Microelectromech. Syst. 2003, 12, 128–137. [CrossRef] Jin, X.C.; Oralkan, O.; Degertekin, F.L.; Khuri-Yakub, B.T. Characterization of one-dimensional capacitive micro-machined ultrasonic immersion transducer arrays. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2001, 48, 750–759. [PubMed] Bayram, B.; Oralkan, O.; Ergun, A.S.; Haeggstrom, E.; Yaralioglu, G.G.; KhuriYakub, B.T. Capacitive micro-machined ultrasonic transducer design for high power transmission. IEEE Trans. UFFC 2005, 52, 326–339. [CrossRef]
Sensors 2016, 16, 814
7.
8. 9. 10. 11.
12. 13. 14.
15. 16. 17.
18.
19.
20.
21.
22. 23. 24. 25. 26.
9 of 9
Villanueva, J.M.; Catunda, S.Y.C.; Tanscheit, R.; Pinto, M.M.S. Wind speed measurement data fusion of phase difference and time-of-flight techniques using ultrasonic transducers. In Proceedings of the 2007 IEEE Instrumentation & Measurement Technology Conference IMTC 2007, Warsaw, Poland, 1–3 May 2007; pp. 1–6. Tong, C.; Figueroa, J.F. Guidelines for the use of ultrasonic noninvasive metering techniques. Flow Meas. Instrum. 2002, 13, 125–142. Lynnworth, L.C.; Liu, Y. Ultrasonic flowmeters: Half-century progress report 1955–2005. Ultrasonics 2006, 44, e1371–e1378. [CrossRef] [PubMed] Van Boxel, J.H.; Sterk, G.; Arens, S.M. Sonic anemometers in aeolian sediment transport research. Geomorphology 2004, 59, 131–147. [CrossRef] Catunda, S.Y.C.; Pessanha, J.E.O.; FonsecaNeto, J.V.; Camelo, N.J.; Silva, P.R.M. Uncertainty analysis for defining a wind power density measurement system structure. In Proceedings of the IEEE Instrumentation and Measurement Technology Conference, Como, Italy, 18–20 May 2004; Volume 2, pp. 1043–1047. Mauricio, J.M.; Catunda, S.Y.C.; Tanscheit, R. Maximum-likelihood data fusion of phase-difference and threshold-detection techniques for wind speed measurement. IEEE Trans. Instrum. Meas. 2009, 58, 2189–2195. Marioli, D.; Narduzzi, C.; Offelli, C.; Petri, D.; Sardini, E.; Taroni, A. Digital time-of-flight measurement for ultrasonic sensor. IEEE Trans. Instrum. Meas. 1992, 41, 93–97. [CrossRef] Espinoza, C.E.M.; Villanueva, J.M.M.; Catunda, S.Y.C. Wind speed measurement and uncertainty using ultrasonic sensor with Kalman filtering. In Proceedings of the Instrumentation and Measurement Technology Conference (I2MTC), Austin, TX, USA, 3–6 May 2010; pp. 624–628. Beck, M.S. Correlation in instruments: Cross correlation flowmeters. J. Phys. E: Sci. Instrum. 1981, 14, 7–19. [CrossRef] Wang, M.; Chen, J. Volumetric Flow Measurement Using an Implantable CMUT Array. IEEE Trans. Biomed. Circ. Syst. 2011, 5, 214–222. [CrossRef] [PubMed] Wang, M.; Chen, J.; Cheng, X.; Zhang, T.; Liu, X. A Bi-directional Real-time Blood Flowmeter Using an Implantable CMUT Array. In Proceedings of the 2008 IEEE Ultrasonics Symposium, Beijing, China, 2–5 November 2008; pp. 1–4. Hoffmann, M.; Unger, A.; Ho, M.C.; Park, K.K.; Khuri-Yakub, B.T.; Kupnik, M. Volumetric characterization of ultrasonic transducers for gas flow metering. In Proceedings of the 2013 IEEE International Ultrasonics Symposium (IUS), Prague, Czech Republic, 21–25 July 2013; pp. 1315–1318. Thranhardt, M.; Mooshofer, P.E.; Hauptmann, P.; Degertekin, L. A Resonant CMUT Sensor for Fluid Applications. In Proceedings of the 2009 IEEE Sensors, Christchurch, New Zealand, 25–28 October 2009; pp. 878–882. Bucci, G.; Ciancetta, F.; Fiorucci, E.; Gallo, D.; Landi, C.; Luiso, M. A Low-Cost Ultrasonic Wind Speed and Direction Measurement System. In Proceedings of the 2013 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Minneapolis, MN, USA, 6–9 May 2013; pp. 505–510. Han, D.; Park, S. A Study on Characteristics of Continuous Wave Ultrasonic Anemometer. In Proceedings of the 2011 IEEE Sensors Applications Symposium (SAS), San Antonio, TX, USA, 22–24 February 2011; pp. 119–122. Han, D.; Kim, S.; Park, S. Two-dimensional ultrasonic anemometer using the directivity angle of an ultrasonic sensor. Microelectronics 2008, 39, 1195–1199. [CrossRef] Raine, A.B.; Aslam, N.; Underwood, C.P.; Danaher, S. Development of an Ultrasonic Airflow Measurement Device for Ducted Air. Sensors 2015, 15, 10705–10722. [CrossRef] [PubMed] Bohn, D.A. Environmental Effects on the Speed of Sound. J. Audio Eng. Soc. 1988, 36, 223–231. Weast, R.C. CRC Handbook of Chemistry and Physics, 67th ed.; CRC Press: Boca Raton, FL, USA, 1986. Lofqvist, T.; Sokas, K.; Delsing, J. Speed of sound measurements in humid air using an ultrasonic flow meter. In Proceedings of the XVII IMEKO World Congress, Dubrovnik, Croatia, 22–27 June 2003; pp. 1650–1653. © 2016 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/).
Two Capacitive Micro-Machined Ultrasonic Transducers for Wind Speed Measurement Gia Thinh Bui, Yu-Tsung Jiang and Da-Chen Pang * Department of Mechanical Engineering, National Kaohsiung University of Applied Sciences, 415 Jiangong Road, Sanmin District, Kaohsiung 807, Taiwan; [email protected] (G.T.B.); [email protected] (Y.J.) * Correspondence: [email protected]; Tel.: +886-7-381-4526 (ext. 5331) Academic Editor: Xiaoning Jiang Received: 7 April 2016; Accepted: 30 May 2016; Published: 2 June 2016
Abstract: This paper presents a new wind speed measurement method using a single capacitive micro-machined ultrasonic transducer (CMUT). The CMUT was arranged perpendicular to the direction of the wind flow, and a reflector was set up a short distance away, facing the CMUT. To reduce the size, weight, cost, and power consumption of conventional ultrasonic anemometers this study proposes two CMUT designs for the measurement of wind speed using either the amplitude of the signal or the time of flight (TOF). Each CMUT with a double array element design can transmit and receive signals in five different operation modes. Experiments showed that the two CMUT designs utilizing the TOF were better than those utilizing the amplitude of the signal for wind speed measurements ranging from 1 m/s to 10 m/s, providing a measurement error of less than 0.2 m/s. These results indicate that the sensitivity of the TOF is independent of the five operation modes. Keywords: capacitive micro-machined ultrasonic transducer (CMUT); wind speed measurement; amplitude; time-of-flight
1. Introduction Capacitive micro-machined ultrasonic transducers (CMUTs) emerged as an interesting alternative to piezoelectric transducers in the mid-1990s [1,2]. Over the past two decades, several research studies have been conducted to explore the fabrication [3,4], characterization [5], and modeling [6] of CMUTs. Our goal in conducting this research was to develop new CMUT array designs that are useful for measuring wind speed through the use of a single CMUT. The measurement of fluid speed, particularly wind speed, using ultrasonic transducers has been found to have important scientific and industrial applications [7,8]. Wind speed measurements can be applied not only in industrial fields, but also for the purpose of environmental monitoring and the control of various systems, such as indoor air conditioning systems, weather forecasting systems, and other sensors [9,10]. Some applications, such as those providing an evaluative measurement of wind speed for wind power generation, require measurements with low levels of uncertainty [11]. To design an instrument for measuring wind speed with measurement uncertainty that is as low as possible, a detailed study of the sources of measurement uncertainties and their propagation was conducted [12]. A common approach for determining wind speed is to measure the time-of-flight (TOF) by means of an ultrasonic signal, where the TOF is defined as the time required for an ultrasonic wave to travel from its source of transmission to the receiving transducer [13]. In the existing literature, several measurement techniques based on TOF estimates have been presented and discussed, including the threshold detection (TH), phase difference (PD), cross-correlation, maximum likelihood estimation (MLE), Kalman filter and wavelet transform techniques [14]. The TH and PD techniques are straightforward when compared with other techniques that make use of signal correlation [15]. However, an estimation of the TOF requires the consideration of factors such as attenuation of Sensors 2016, 16, 814; doi:10.3390/s16060814
www.mdpi.com/journal/sensors
Sensors 2016, 16, 814
2 of 9
the medium, random noise, and reflections, which lower the signal noise ratio (SNR) and increase uncertainty to high levels. To develop a minimally complex CMUT array for ultrasound transmissions and reception, some publications recommended an ultrasound flow measurement system with a transmitter and receiver on the same substrate as our system. Instead of using Doppler effects for flow measurement [16,17], we used the TOF method to determine wind velocity. In the comparison with the methods used in previous studies [18,19], our method also has the advantage of allowing for the quick calculation of results from simple equations. In this paper, we propose two CMUT designs to measure wind speed, based on the TOF and amplitude of ultrasonic signals between transmission and reception. Specifically, we use only a single CMUT to yield a measurement of velocity with small spatial resolution of 10 mm between the reflector and transducer. The findings from this study can reduce the number of sensors needed, compared with previous systems [20–22]. Furthermore, the 3 mm ˆ 3 mm CMUT presented in this study is also smaller than other ultrasonic anemometers that have been used previously [23]. Because of the smaller distance between the CMUT and reflector, the travel time is shorter and the CMUT’s sampling rate is higher. The CMUTs for wind speed measurement have a novel design for the transmission and reception of signals in five different modes of operation. This approach also uses the available array element area efficiently. In this paper, the implementation of two new CMUT designs for wind speed testing is described, and the pulse-echo testing results are discussed. These results show that a single CMUT can be used for wind speed measurement with good accuracy. In addition, the whole system can be realized with a simple structure and low costs. 2. CMUT Design The basic mechanical structure of a CMUT is similar to that of a capacitor with two parallel plates (see Figure 1). The membrane, filling layer, and side wall are all made of polymer, an epoxy-type photoresist (SU-8). The filling layer is designed to protect the gold top electrode from the ambience and external corrosion during regular operation. A thin indium tin oxide bottom electrode is adhered on a substrate and separated from the membrane by an air cavity. The geometric parameters of a CMUT are summarized Sensors 2016, 16, 814 in Table 1. 3 of 10
Figure 1. Cross-sectional schematic of a single CMUT cell. Figure 1. Cross-sectional schematic of a single CMUT cell. Table 1. Mechanical dimensions of a CMUT cell. Table 1. Mechanical dimensions of a CMUT cell. Parameter Measurement (μm) Membrane diameter 140 Parameter Measurement (µm) Membrane thickness 5 Membrane diameter 140 Sidewall width 105 Membrane thickness Cavitywidth height 210 Sidewall Bottom electrode 0.05 Cavity heightthickness 2 Bottom electrode thickness 0.05 Top electrode diameter 108 Top diameter 108 Topelectrode electrode thickness 0.15 Top electrode thickness
0.15
Sensors 2016, 16, 814
3 of 9
Figure 1. Cross-sectional schematic of a single CMUT cell.
The aim of this research is to design new CMUTs with of divergent characteristics. The transmission Figure 1. Cross-sectional schematic a single CMUT cell. and reception modes of a single device can be operated separately so that the element can be adjusted Table 1. Mechanical dimensions of a CMUT cell. 1. ultrasonic Mechanicalsignals. dimensions a CMUT cell.as designed can work with individually to transmit andTable receive The of single device Parameter Measurement (μm) five operation modes using different ratio of transmission and reception cells. In this study, two Parameter Measurement (μm) Membrane 140 CMUT designs are presented, as showndiameter in Figure 2. Each one was arranged in a square area measuring Membrane diameter 140 Membrane thicknessof 140 and 5 µm, respectively, 5 3 mm ˆ 3 mm with a diameter and thickness for each cell membrane Membrane thickness 5 Sidewall width 10 and a cavity height of 2 µm. Each CMUT has two arrays of cells, called the majority and minority Sidewall width 10 Cavity height 2 arrays. The majority array has more cells than the minority array, with2a 2:1 ratio, in these designs. In Cavity height Bottom electrode thickness 0.05 the wave-type CMUT design, the array element is separated using 0.1 mm spacing to include 416 cells Bottom electrode thickness 0.05 Top electrode diameter 108 inside it, as shown in Figure Top 2a. Figure 2b shows the parallel-line type CMUT arranged in a rectangular electrode diameter 108 Top electrode thickness 0.15 array, with a majority arrayTop of 80 cells arranged electrode thicknessin a 3 mm ˆ 0.65 mm 0.15area and a minority array of 40 cells arranged in a 3 mm ˆ 0.35 mm area.
(a) (a)
(b) (b)
Figure 2. Diagram of two CMUT designs: (a) Wave-type; (b) Parallel-line type. Figure of two twoCMUT CMUTdesigns: designs:(a) (a)Wave-type; Wave-type; Parallel-line type. Figure 2. 2. Diagram Diagram of (b)(b) Parallel-line type.
3. Measurement Principle 3.3.Measurement MeasurementPrinciple Principle It is common to use two ultrasonic transducers for measuring wind speed by lining up the ItItisis common to use two two ultrasonic ultrasonictransducers transducersfor formeasuring measuring wind speed lining common to use speed by by lining up up the the transmitting and receiving transducers at a specific angle. This study wind proposes a new measurement transmitting and receiving transducers at a specific angle. This study proposes a new measurement transmitting and receiving transducers at a specific angle. This study proposes a new measurement system that uses only one CMUT and one reflector within a small space as shown in Figure 3. In this system CMUT and andone onereflector reflectorwithin within a small space shown in Figure In this systemthat thatuses uses only only one CMUT a small space as as shown in Figure 3. In3.this design, the wind flows perpendicular to the ultrasound signal path between the CMUT and the design, perpendicular to tothe theultrasound ultrasoundsignal signalpath path between CMUT design,the the wind wind flows flows perpendicular between thethe CMUT andand the the reflector. The wind will change the ultrasound propagation by increasing the TOF and reducing the reflector.The The wind wind will change increasing thethe TOF andand reducing the the reflector. change the theultrasound ultrasoundpropagation propagationbyby increasing TOF reducing amplitude of the ultrasound signals. As shown in Figure 3, the red vector is the ultrasonic path when amplitudeof ofthe the ultrasound ultrasound signals. red vector is the ultrasonic path when amplitude signals.As Asshown shownininFigure Figure3,3,the the red vector is the ultrasonic path when it is altered by the wind flow, d is the distance (i.e., 10 mm) between the CMUT and the reflector, and alteredby bythe the wind wind flow, flow, dd is thethe CMUT and thethe reflector, andand ititisisaltered isthe thedistance distance(i.e., (i.e.,1010mm) mm)between between CMUT and reflector, θ isθ the angle between the transmitted path of the ultrasonic signal and andthe theshifted shiftedcondition conditionofof the is the angle between the transmitted path of the ultrasonic signal θ is the angle between the transmitted path of the ultrasonic signal and the shifted conditionthe of the wind direction. winddirection. direction. wind
Figure the CMUT. CMUT. Figure3.3.Wind Windspeed speedmeasurement measurement using the Figure 3. Wind speed measurement using the CMUT.
3.1. Measuring Wind Velocity In this study, a method of measuring wind velocity was developed based on the different measurements times of the transmission time of the ultrasonic signals. This measured time difference ∆t is expressed as Equation (6), which is derived from Equations (4) and (5). The latter equations are
measurements times of the transmission time of the ultrasonic signals. This measured time difference Δt is expressed as Equation (6), which is derived from Equations (4) and (5). The latter equations are themselves derived from Equations (2) and (3) in terms of the velocity of wind v being much lower than sound speed c. The time difference of the ultrasonic signals transmitted Δt1 and echoed Δt2 were determined by substituting the measured transmitted time t1 and echo time t2. The Sensors 2016, 16, 814 4 of 9 transmission time of the ultrasonic signal t0 to the reflector without wind is measured as presented in Equation (1): themselves derived from Equations (2) and (3) in terms of the velocity of wind v being much lower d t than sound speed c. The time difference of the ultrasonic signals transmitted ∆t1 and echoed ∆t2 were(1) c determined by substituting the measured transmitted time t1 and echo time t2 . The transmission time of the ultrasonic signal t0 to the reflector without wind is measured as presented in Equation (1):
t
d
vt
t
c
1
v c
1 v d 2 c t0 “ t
t
1 v 2 c
t
⋯
1 v 2 c
t
c
tt1 “ b
b dd2 ` pvt 2vt0 q2 cc
d2 ` p2vt0 q2
c ´ v2v ¯2 ´1 ¯ ´1 ¯2v ´1 ¯2v 2v 2 4 n t t` 1 v t0 ` t 1 v t0 `t . . . ` ⋯1 v t0 t “ t0 11` “ t 0 cc 2 2c c 2 2c c 2 2c c
(3)
(2)
d
˙ ˆ ˙ ˆ ˙ ˆ ˙ 1 2v1 v2 1 2v 4 1 2v n 2v 2 t0 ` t0 ` . . . ` t0 t ∆t “ t0 ` t c c 2 c2 c 2 c 2 c 1 ´ v ¯2 ∆t1 “ t1 ´ t0 « 1 2v t0 t t 2 c t ∆t ˆ2 c˙2 1 2v ∆t2 “ t2 ´ t0 « t0 2 5c v ∆t ∆t ∆t c2 d 5 ´2v ¯ ∆t “ ∆t1 ` ∆t2 “ Temperature and humidity affect the density of the 2 cair, ccausing changes in the velocity t2 “
(2) (1)
ˆ
“ t0
1`
(3)(4) (4)
(5)
(5)
(6) (6)
of the ultrasound signal [24–26]. If the CMUT used at room temperature less than 1 °C variation, Temperature and humidity affectisthe density of the air, causingwith changes in the velocity of thethe ˝ ultrasound velocity change is within 0.2%. The humidity effect on ultrasound velocity is the even ultrasound signal [24–26]. If the CMUT is used at room temperature with less than 1 C variation, smaller. The ultrasound velocity change only 0.05% ifeffect the temperature constant and humidity ultrasound velocity change is within 0.2%.isThe humidity on ultrasoundisvelocity is even smaller. The ultrasound velocity change is only 0.05% if the temperature is constant and humidity changes changes remain within 10%. remain within 10%.
3.2. Wind Speed Measurement System Configuration 3.2. Wind Speed Measurement System Configuration
A wind speed measurement system was initiated for basic measurements, as shown in Figure 4. A wind speed measurement system was initiated for basic measurements, as shown in Figure 4. The measurement box had the dimensions of 120 cm × 60 cm × 60 cm. On cross-sections The measurement box had the dimensions of 120 cm ˆ 60 cm ˆ 60 cm. On cross-sections perpendicular perpendicular to the wind flow direction, the wind velocity was determined on a grid. A transducer to the wind flow direction, the wind velocity was determined on a grid. A transducer was installed on was installed on the calibration system. The wind speed tests were performed in the laboratory at a the calibration system. The wind speed tests were performed in the laboratory at a temperature of temperature of 25 °C ± 0.5 °C and humidity of 60% ± 3%. The effects of temperature and humidity on 25 ˝ C ˘ 0.5 ˝ C and humidity of 60% ˘ 3%. The effects of temperature and humidity on ultrasound ultrasound velocity were very small. The environment uncertainty was approximately 0.2%. velocity were very small. The environment uncertainty was approximately 0.2%.
Figure4.4.Photograph Photographof ofthe thewind wind speed measurement Figure measurementconfiguration. configuration.
Figure 5 shows the setup for the wind speed experiment. The CMUT was powered by DC 100 V and AC 300 V power sources, which produced the ultrasonic pulse signals used in the experiment. A generator (DPR 300 Pulse/Receiver, JSR Ultrasonics, Pittsford, NY, USA) transmits a short pulse, which has an amplitude of 0–475 V. A variable DC bias voltage of 0–300 V (N5751A DC Power Supply, Keysight Technologies, Santa Rosa, CA, USA) is superimposed on the transient voltage. The receiver signals were obtained with an oscilloscope (DPO 4104B Digital Phosphor Oscilloscope, Tektronix,
Figure 5 shows the setup for the wind speed experiment. The CMUT was powered by DC 100 V Sensors 2016, 16, 814 5 of 10 and AC 300 V power sources, which produced the ultrasonic pulse signals used in the experiment. A generator (DPR 300 Pulse/Receiver, Ultrasonics, Pittsford, USA) a DC short Figure 5 shows the setup for theJSR wind speed experiment. The NY, CMUT was transmits powered by 100pulse, V which an V amplitude of 0–475 V.produced A variable bias voltage of 0–300 DC Power and has AC 300 power sources, which the DC ultrasonic pulse signals usedV in (N5751A the experiment. A Sensors 2016, 16, 814 5 of 9 generator (DPRTechnologies, 300 Pulse/Receiver, Ultrasonics, NY, USA) a short pulse,The Supply, Keysight Santa JSR Rosa, CA, USA)Pittsford, is superimposed ontransmits the transient voltage. whichsignals has an were amplitude of 0–475 A variable DC bias voltage 0–300 VPhosphor (N5751A DC Power receiver obtained withV.an oscilloscope (DPO 4104Bof Digital Oscilloscope, Supply, Keysight Technologies, Santa Rosa, CA, USA) is superimposed on the transient voltage. The Tektronix, OR, Beaverton, OR, USA)toconnected the TOFs CMUT. Any TOFs in the measurements Beaverton, USA) connected the CMUT.toAny obtained in theobtained measurements were recorded receiver signals were obtained with an oscilloscope (DPO 4104B Digital Phosphor Oscilloscope, weretransmitted recorded and and to atransmitted computer. to a computer. Tektronix, Beaverton, OR, USA) connected to the CMUT. Any TOFs obtained in the measurements were recorded and transmitted to a computer. Receiver output T/R
Receiver output
TRIG/SYNC
Through
T/R
Pulse/Receiver Through Generator
Oscilloscope
TRIG/SYNC
Pulse/Receiver Generator
Bias tee
DC Power Supply
Bias tee
DC Power Supply
Oscilloscope
Ch1
CMUT (Transmitter) CMUT (Transmitter)
CMUT (Receiver)
Ch1
EXT/TRIG
EXT/TRIG
Bias tee
CMUT Bias (Receiver) tee Ultrasound
Reflector
DC Power Supply DC Power Supply
Ultrasound
Reflector
5. Setup for wind speed experiment. Figure 5. Figure 5. Setup for wind speed experiment.
Figure 66 shows shows the the received received ultrasonic signal’s signal’s amplitude. The The measurement measurement system system starts starts its its Figure Figure 6 shows the receivedultrasonic ultrasonic signal’s amplitude. amplitude. The measurement system starts its operation with a pulse-wave transmitted and received by the transducer, which is propagated in the operation with a pulse-wave bythe thetransducer, transducer,which which is propagated in the operation with a pulse-wavetransmitted transmittedand and received received by is propagated in the medium. The time needed for the detection of the received ultrasonic signal was determined using a medium. The time needed ofthe thereceived receivedultrasonic ultrasonic signal determined medium. The time neededfor forthe thedetection detection of signal waswas determined usingusing a reference threshold. From this detection, a data acquisition process utilizing the transmitted and a reference dataacquisition acquisitionprocess processutilizing utilizing transmitted referencethreshold. threshold. From From this detection, detection, aa data thethe transmitted andand received signals was conducted to determine any and all detection times. Then, an analysis of the the received signals was conductedtotodetermine determine any any and all anan analysis of the received signals was conducted all detection detectiontimes. times.Then, Then, analysis of signals ledled to estimates estimates ofof the PD between the transmitted andreceived receivedsignals. signals. The number of cells cells signals to estimates thePD PDbetween betweenthe the transmitted transmitted and The number of cells signals led to of the and received signals. The number of and the signal amplitudes of the two CMUT designs in five operation modes are shown in Table signal amplitudes thetwo twoCMUT CMUTdesigns designs in in five shown in in Table 2. 2.2. andand the the signal amplitudes ofofthe five operation operationmodes modesare are shown Table
Figure Timeresponse responseof ofpulse-echo pulse-echo signal mode. Figure 6. 6. Time signalin infull/full full/full mode.
Figure 6. Time response of pulse-echo signal in full/full mode. The operation of the transducer is based on the ratio of transmission and reception cells. The Table 2. Number of cells and signal amplitudes of the two CMUT designs in the five operation modes. experimental measurements have five modes: full transmission/full reception, majority The operation of the transducer is based on the ratio of transmission and reception cells. The transmission/ majority reception, majority reception, minority Wave-Typetransmission/minority Design Parallel-Line Type Design
experimental measurements have five modes: full transmission/full reception, majority Mode Signal Transmission/Reception Signal transmission/ majority reception,Transmission/Reception majority transmission/minority reception, minority (Cell Number) (mV) (Cell Number) (mV) Full transmission/Full reception Majority transmission/Majority reception Majority transmission/Minority reception Minority transmission/Majority reception Minority transmission/Minority reception
416/416 274/274 274/142 142/274 142/142
634 362 226 244 158
360/360 240/240 240/120 120/240 120/120
546 324 196 214 140
transmission/full reception mode had the maximum received signals and the highest sensitivity. The minority transmission/majority reception mode, had fewer transmitted signals than the majority transmission/majority reception mode but still received some signals and was considered better and relatively more efficient. Sensors 2016, 16, 814
Table 2. Number of cells and signal amplitudes of the two CMUT designs in the five operation modes.
6 of 9
The operation of the transducer is basedWave-Type on the ratio of transmission and reception cells. The Design Parallel-Line Type Design experimental measurements have five modes: full transmission/full reception, majority transmission/ Mode Transmission/Reception Signal Transmission/Reception Signal (Cell Number) reception, (mV)minority (Cell Number) (mV) majority reception, majority transmission/minority transmission/majority Full transmission/Full reception 416/416 634 360/360 546 reception, and minority transmission/minority reception. The full transmission/full reception mode Majority transmission/Majority reception 274/274 362 240/240 324 had the maximum received signals and the highest sensitivity. The minority transmission/majority Majority transmission/Minority reception 274/142 226 240/120 196 reception mode, had fewerreception transmitted signals than the majority reception Minority transmission/Majority 142/274 244 transmission/majority 120/240 214 Minority 142/142 158 relatively 120/120 mode buttransmission/Minority still received somereception signals and was considered better and more efficient. 140 4. ExperimentalResults Results 4. Experimental ˝ In this experiment the transducer has a 90 90° directivity angle and a reflecting surface set at a distance of 10 mm. The maximum signal of the first echo was used as an observation point, whereby it was found that the attenuation of the signal increases as the wind wind speed speed increases. increases. The peak time increases slowly along with the increasing wind speed, but the operation time of the first echo and of the resonance took advantage of the time difference between the resonancefrequency frequencydoes doesnot notchange. change.This Thisstudy study took advantage of the time difference between increasing attenuation of the of echo determine wind speed changes. shown inAs Figure 7, the the increasing attenuation thesignals echo to signals to determine wind speedAschanges. shown in experiments measured wind speeds from 1 m/s to 10 m/s and the results of the full transmission/full Figure 7, the experiments measured wind speeds from 1 m/s to 10 m/s and the results of the full reception mode and minoritymode transmission/majority reception showed thatreception the valuesshowed of the maximum transmission/full reception and minority transmission/majority that the points signal points changeofwhen the wind increases. values of of the the echo maximum the echo signalspeed change when the wind speed increases.
(a)
(b)
Figure 7. 7. Increasing to 10 m/s: (a) The full Figure Increasing attenuation attenuationof ofthe theecho echosignal signalby bywind windspeeds speedsfrom from1 m/s 1 m/s to 10 m/s: (a) The transmission/full reception modemode results; (b) The transmission/majority reception mode full transmission/full reception results; (b) minority The minority transmission/majority reception results. mode results.
4.1. Using Amplitude of Signal to Measure Wind Speed 4.1. Using Amplitude of Signal to Measure Wind Speed As shown in Figure 8a, the wave-type transducer design was tested. The experimental results of As shown in Figure 8a, the wave-type transducer design was tested. The experimental results the five modes were affected by the wind speed, with sensitivity of magnitude signals and a of the five modes were affected by the wind speed, with sensitivity of magnitude signals and a standard deviation of error as shown in Table 3. Figure 8b shows the results achieved when using standard deviation of error as shown in Table 3. Figure 8b shows the results achieved when using the parallel-line type ultrasonic transducer design for measurement. The measurement results of the the parallel-line type ultrasonic transducer design for measurement. The measurement results of the five modes were affected by the wind speed, with attenuation sensitivity of magnitude signals and a five modes were affected by the wind speed, with attenuation sensitivity of magnitude signals and a standard deviation of error as shown in Table 3. The results were based on the least square method standard deviation of error as shown in Table 3. The results were based on the least square method to to draw the best fit line. From the results, it can be concluded that the strength of the wind speed had draw the best fit line. From the results, it can be concluded that the strength of the wind speed had a a direct effect on the amplitude of the ultrasonic signals. The ultrasonic echo signals of the two types direct effect on the amplitude of the ultrasonic signals. The ultrasonic echo signals of the two types of of transducers were smaller when the attenuation of the wind speed increased. The measurements of transducers were smaller when the attenuation of the wind speed increased. The measurements of the two transducers in full transmission/full reception mode demonstrated high sensitivity significance. It was found that the minority transmission/majority reception mode had a larger signal than in majority transmission/minority reception mode.
Sensors 2016, 16, 814
7 of 10
the two transducers in full transmission/full reception mode demonstrated high sensitivity significance. It was found that the minority transmission/majority reception mode had a larger Sensors 2016, 16, 814 7 of 9 signal than in majority transmission/minority reception mode.
(a)
(b)
Figure 8. Effect of wind speed on measurement results: (a) Wave-type CMUT design; (b) Parallel-line Figure 8. Effect of wind speed on measurement results: (a) Wave-type CMUT design; (b) Parallel-line type CMUT design. type CMUT design. Table of measurement measurement for for the the two two CMUT CMUT designs. designs. Table 3. 3. The The sensitivities sensitivities and and standard standard deviation deviation errors errors of Designs Designs
Wave-type CMUT Wave-type CMUT
Parallel-line Parallel-line type type CMUT CMUT
Modes Modes Full/Full Full/Full Majority/Majority Majority/Majority Majority/Minority Majority/Minority Minority/Majority Minority/Majority Minority/Minority Minority/Minority Full/Full Full/Full Majority/Majority Majority/Majority Majority/Minority Majority/Minority Minority/Majority Minority/Majority Minority/Minority Minority/Minority
Amplitude TOF Amplitude TOF Sensitivity Sensitivity Error (m/s) Sensitivity−1 2 Error (m/s) Sensitivity Error Error (mV/ms−1) (ns/(ms ) ) (mV/ms´1 ) (m/s) (ns/(ms´1 )2 ) (m/s) 8.807 0.198 0.617 0.153 8.807 0.198 0.617 0.153 8.379 0.254 0.616 0.154 8.379 0.254 0.616 0.154 7.161 0.235 0.614 0.168 7.161 0.235 0.614 0.168 7.555 0.272 0.617 0.161 7.555 0.272 0.617 0.161 6.846 0.330 0.616 0.165 6.846 0.330 0.616 0.165 9.906 0.189 0.618 0.155 9.906 0.189 0.618 0.155 9.552 0.180 0.616 0.156 9.552 0.180 0.616 0.156 8.944 0.249 0.615 0.170 8.944 0.249 0.615 0.170 9.166 0.241 0.617 0.165 9.166 0.241 0.617 0.165 8.171 0.288 0.617 0.172 8.171 0.288 0.617 0.172
4.2. Using Differences in TOF’s to Measure Wind Speed 4.2. Using Differences in TOF’s to Measure Wind Speed In this study, the transducers used self-transmitted and self-received signals sent over a reflector In this study, the transducers used self-transmitted and self-received signals sent over a distance of 10 mm to measure wind speed. The influence of the wind causes the point of a signal’s reflector distance of 10 mm to measure wind speed. The influence of the wind causes the point of a arrival on the reflector to be offset by a distance of vt, and the wind effect on the speed of an ultrasound signal’s arrival on the reflector to be offset by a distance of vt, and the wind effect on the speed of an signal is offset again after reflection as shown in Figure 3. The transducers were used in experiments ultrasound signal is offset again after reflection as shown in Figure 3. The transducers were used in to measure wind speeds ranging from 1 m/s to 10 m/s. With greater wind speeds, the time required experiments to measure wind speeds ranging from 1 m/s to 10 m/s. With greater wind speeds, the to receive signals increases. The TOF can be calculated from the experimental results and compared time required to receive signals increases. The TOF can be calculated from the experimental results with the 2016, theoretical 16, 814 results as shown in Figure 9. 8 of 10 andSensors compared with the theoretical results as shown in Figure 9. There are significant errors evident when the wind speed is small. This contradicts the presented theoretical values. The results utilized the least square method to draw the second order curve fitting. Table 3 shows the wind speed measurement error using a difference in TOF that is slightly lower than the amplitude of the signal.
(a)
(b)
Figure 9. Comparison of the results from the theoretical calculation and the experimental Figure 9. Comparison of the results from the theoretical calculation and the experimental measurements measurements according to different TOFs: (a) Wave-type CMUT design; (b) Parallel-line type according to different TOFs: (a) Wave-type CMUT design; (b) Parallel-line type CMUT design. CMUT design.
5. Conclusions In this paper, two designs of a CMUT with five modes that can be used to measure the velocity of wind blowing perpendicular to the CMUT were presented. In addition, analyses of the ultrasound signal amplitudes and the TOF differences due to different wind speeds were conducted. The
Sensors 2016, 16, 814
8 of 9
There are significant errors evident when the wind speed is small. This contradicts the presented theoretical values. The results utilized the least square method to draw the second order curve fitting. Table 3 shows the wind speed measurement error using a difference in TOF that is slightly lower than the amplitude of the signal. 5. Conclusions In this paper, two designs of a CMUT with five modes that can be used to measure the velocity of wind blowing perpendicular to the CMUT were presented. In addition, analyses of the ultrasound signal amplitudes and the TOF differences due to different wind speeds were conducted. The problems with the measurement range for the TOF differences and signal amplitudes of the continuous-wave CMUT were compared. The two CMUT designs that used the TOF method did not show much of a difference in terms of sensitivity and error. The TOF determination method can be used to measure wind speeds without calibration. Also, it was observed that the full transmission/full reception mode was the best choice in terms of providing greater sensitivity and less uncertainty in wind speed measurement when using the amplitude of signal method. The minority transmission/majority reception mode is proposed for good energy efficiency since the design can be selected for signal transmission. The parallel-line type design produced results that were only slightly better than those of the wave-type design. Thus, the two CMUT designs can be used to build a wind speed measurement system that is independent of the five modes and has the advantages of high accuracy, low cost, and easy portability, as well as great potential for future applications. Acknowledgments: The authors would like to acknowledge the financial support of MEMS and Precision Machinery Research and Development Center at National Kaohsiung University of Applied Sciences for the development of the CMUTs. We would also like to acknowledge the Ministry of Sciences and Technology (MOST) for its partial support under project number MOST 104-2221-E-151-012. The wind speed experiment was performed at the Thermo-Fluid-Optical Visualization laboratory of the Department of Mold and Die Engineering at National Kaohsiung University of Applied Sciences. Author Contributions: Da Chen Pang and Yu Tsung Jiang conceived and designed the experiments; Yu Tsung Jiang performed the experiments; Da Chen Pang and Gia Thinh Bui analyzed the data; Gia Thinh Bui and Da Chen Pang wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.
References 1. 2.
3.
4. 5.
6.
Haller, M.I.; Khuri-Yakub, B.T. A surface micro-machined electrostatic ultrasonic air transducer. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 1996, 43, 1–6. [CrossRef] Ladabaum, I.; Khuri-Yakub, B.T.; Spoliansky, D.; Haller, M.I. Micro-machined ultrasonic transducers (MUTs). In Proceedings of the 1995 IEEE Ultrasonics Symposium, Seattle, WA, USA, 7–10 November 1995; Volume 1, pp. 501–504. Jin, X.C.; Ladabaum, I.; Degertekin, F.L.; Calmes, S.; KhuriYakub, B.T. Fabrication and characterization of surface surface micromachined capacitive ultrasonic immersion transducers. J. Microelectromech. Syst. 1999, 8, 100–114. Huang, Y.; Ergun, A.S.; Haeggstrom, E.; Badi, M.H.; KhuriYakub, B.T. Fabricating capacitive micro-machined ultrasonic transducers with wafer-bonding. J. Microelectromech. Syst. 2003, 12, 128–137. [CrossRef] Jin, X.C.; Oralkan, O.; Degertekin, F.L.; Khuri-Yakub, B.T. Characterization of one-dimensional capacitive micro-machined ultrasonic immersion transducer arrays. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2001, 48, 750–759. [PubMed] Bayram, B.; Oralkan, O.; Ergun, A.S.; Haeggstrom, E.; Yaralioglu, G.G.; KhuriYakub, B.T. Capacitive micro-machined ultrasonic transducer design for high power transmission. IEEE Trans. UFFC 2005, 52, 326–339. [CrossRef]
Sensors 2016, 16, 814
7.
8. 9. 10. 11.
12. 13. 14.
15. 16. 17.
18.
19.
20.
21.
22. 23. 24. 25. 26.
9 of 9
Villanueva, J.M.; Catunda, S.Y.C.; Tanscheit, R.; Pinto, M.M.S. Wind speed measurement data fusion of phase difference and time-of-flight techniques using ultrasonic transducers. In Proceedings of the 2007 IEEE Instrumentation & Measurement Technology Conference IMTC 2007, Warsaw, Poland, 1–3 May 2007; pp. 1–6. Tong, C.; Figueroa, J.F. Guidelines for the use of ultrasonic noninvasive metering techniques. Flow Meas. Instrum. 2002, 13, 125–142. Lynnworth, L.C.; Liu, Y. Ultrasonic flowmeters: Half-century progress report 1955–2005. Ultrasonics 2006, 44, e1371–e1378. [CrossRef] [PubMed] Van Boxel, J.H.; Sterk, G.; Arens, S.M. Sonic anemometers in aeolian sediment transport research. Geomorphology 2004, 59, 131–147. [CrossRef] Catunda, S.Y.C.; Pessanha, J.E.O.; FonsecaNeto, J.V.; Camelo, N.J.; Silva, P.R.M. Uncertainty analysis for defining a wind power density measurement system structure. In Proceedings of the IEEE Instrumentation and Measurement Technology Conference, Como, Italy, 18–20 May 2004; Volume 2, pp. 1043–1047. Mauricio, J.M.; Catunda, S.Y.C.; Tanscheit, R. Maximum-likelihood data fusion of phase-difference and threshold-detection techniques for wind speed measurement. IEEE Trans. Instrum. Meas. 2009, 58, 2189–2195. Marioli, D.; Narduzzi, C.; Offelli, C.; Petri, D.; Sardini, E.; Taroni, A. Digital time-of-flight measurement for ultrasonic sensor. IEEE Trans. Instrum. Meas. 1992, 41, 93–97. [CrossRef] Espinoza, C.E.M.; Villanueva, J.M.M.; Catunda, S.Y.C. Wind speed measurement and uncertainty using ultrasonic sensor with Kalman filtering. In Proceedings of the Instrumentation and Measurement Technology Conference (I2MTC), Austin, TX, USA, 3–6 May 2010; pp. 624–628. Beck, M.S. Correlation in instruments: Cross correlation flowmeters. J. Phys. E: Sci. Instrum. 1981, 14, 7–19. [CrossRef] Wang, M.; Chen, J. Volumetric Flow Measurement Using an Implantable CMUT Array. IEEE Trans. Biomed. Circ. Syst. 2011, 5, 214–222. [CrossRef] [PubMed] Wang, M.; Chen, J.; Cheng, X.; Zhang, T.; Liu, X. A Bi-directional Real-time Blood Flowmeter Using an Implantable CMUT Array. In Proceedings of the 2008 IEEE Ultrasonics Symposium, Beijing, China, 2–5 November 2008; pp. 1–4. Hoffmann, M.; Unger, A.; Ho, M.C.; Park, K.K.; Khuri-Yakub, B.T.; Kupnik, M. Volumetric characterization of ultrasonic transducers for gas flow metering. In Proceedings of the 2013 IEEE International Ultrasonics Symposium (IUS), Prague, Czech Republic, 21–25 July 2013; pp. 1315–1318. Thranhardt, M.; Mooshofer, P.E.; Hauptmann, P.; Degertekin, L. A Resonant CMUT Sensor for Fluid Applications. In Proceedings of the 2009 IEEE Sensors, Christchurch, New Zealand, 25–28 October 2009; pp. 878–882. Bucci, G.; Ciancetta, F.; Fiorucci, E.; Gallo, D.; Landi, C.; Luiso, M. A Low-Cost Ultrasonic Wind Speed and Direction Measurement System. In Proceedings of the 2013 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Minneapolis, MN, USA, 6–9 May 2013; pp. 505–510. Han, D.; Park, S. A Study on Characteristics of Continuous Wave Ultrasonic Anemometer. In Proceedings of the 2011 IEEE Sensors Applications Symposium (SAS), San Antonio, TX, USA, 22–24 February 2011; pp. 119–122. Han, D.; Kim, S.; Park, S. Two-dimensional ultrasonic anemometer using the directivity angle of an ultrasonic sensor. Microelectronics 2008, 39, 1195–1199. [CrossRef] Raine, A.B.; Aslam, N.; Underwood, C.P.; Danaher, S. Development of an Ultrasonic Airflow Measurement Device for Ducted Air. Sensors 2015, 15, 10705–10722. [CrossRef] [PubMed] Bohn, D.A. Environmental Effects on the Speed of Sound. J. Audio Eng. Soc. 1988, 36, 223–231. Weast, R.C. CRC Handbook of Chemistry and Physics, 67th ed.; CRC Press: Boca Raton, FL, USA, 1986. Lofqvist, T.; Sokas, K.; Delsing, J. Speed of sound measurements in humid air using an ultrasonic flow meter. In Proceedings of the XVII IMEKO World Congress, Dubrovnik, Croatia, 22–27 June 2003; pp. 1650–1653. © 2016 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/).
