A novel, high resolution temperature sensor for balloon applications

A novel, high resolution temperature sensor for balloon applications

~ Pergamon www.elsevier.com/locate/asr Adv. Space Res. Vol. 30, No. 5, pp. 1365-1369, 2002 © 2002 COSPAR. Published by Elsevier Science Ltd. All ri...

313KB Sizes 1 Downloads 49 Views

~

Pergamon

www.elsevier.com/locate/asr

Adv. Space Res. Vol. 30, No. 5, pp. 1365-1369, 2002 © 2002 COSPAR. Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0273-1177/02 $22.00 + 0.00 PII: S0273-1177(02)00552-5

A NOVEL, HIGH RESOLUTION TEMPERATURE SENSOR FOR BALLOON APPLICATIONS M. Friedrich ~, M. Posch 1'2'3, S. Kirkwood 2, K. Stebel z, and K. Torkar 4

JDepartment of Communications and Wave Propagation, Technical University Graz, Inffeldgasse 12, A-8010 Graz, Austria 2MRI Atmospheric Research Programme, Swedish Institute of Space Physics, Kiruna, Sweden L~nowwith: Philips Semiconductors, Gratkorn, Austria 4Space Research Institute, Austrian Academy of Sciences, Schmiedlstrasse 6, A-8042 Graz, Austria ABSTRACT The speed of sound is to a very good approximation proportional to the square root of the temperature. A temperature sensor based on that principle was flown from ESRANGE, Sweden, in September 1999 together with a temperature sensor of an ozone sonde. The new sensor provided data comparable to the conventional sensor up to the apogee of 19 km of that particular balloon flight. The high-resolution variations of the acoustically derived temperatures are analysed relative to the radiosonde data, but also as frequency spectra. Explanations for differences from the theoretical behaviour are put forward and conceivable improvements are presented. © 2002 COSPAR. Published by Elsevier Science Ltd. All rights reserved.

INTRODUCTION At balloon altitudes (<40 km) the air's density is large enough to use temperature sensors that rely on complete thermal equilibrium between the ambient air and the actual sensor (wire). Generations of meteorologists have developed the necessary corrections to account for the influence of direct solar radiation or aerodynamic effects. For conventional meteorological purposes one measurement per second corresponding typically to about one reading every 5 metres, is more than adequate. There are, however, indications that the temperature profile in the tropoand stratosphere may contain more details than suggested by the monotonous decrease to the tropopause, followed by an increase to the ozone maximum in the stratosphere. One such phenomenon is the reflection of radio waves in the VHF range, a fact that is made use of in MST (mesosphere, stratosphere, troposphere) radars; on the other hand any oblique radar can be disturbed by such "clear air" disturbances. The relation to MST radar echoes stimulated the development of a high resolution thermometer. Radar theory requires the refractive index to change appreciably within half the radar wavelength in order to produce an echo. At balloon altitudes the atmosphere is not ionised, hence a change in the refractive index can only be due to a variation of the dielectric constant. If one can rule out drastic changes in the composition of the major constituents (which one can safely do in the tropo- and stratosphere), only changes in the temperature can produce the observed effect (radar echoes). In order to qualitatively check MST radar theories a balloon borne instrument was developed which provides the necessary temperature and time (altitude) resolution. At least one such dedicated measurement campaign has been conducted (Luce et al., 1995) who used two vertically spaced balloon borne sensors, each essentially of conventional type, together with observations of humidity and velocity. Since the prime purpose of the flight to be described here was to test the potential of a new sensor no such elaborate set of measurements was provided. INSTRUMENT The speed of sound c is to a very good approximation proportional to ~ provided the composition of the major constituents is known, which certainly is the case for balloon altitudes. A more rigorous relation also includes humidity which can conveniently be expressed as partial water vapour pressure e relative to the total pressure p: 1365

M. Friedrich et al.

1366

c= Ix--~(l

+O.32 e ) T

=

(1)

ms-'

K = cp/c,, = 1.402 for di-atomic gases, R = 8.3145 J K-lmol -~, M = 28.9647 kg kmol -~ There are basically two ways to exploit this relation for the determination of temperature (a) to measure the time it takes a tone burst to travel over a known distance d, or (b) to measure the phase of a CW signal; we employ the latter method. In short, we use a central loudspeaker and two horizontal propagation paths in opposite directions of 0.5 m each. In each path the phase between a near and distant microphone is measured in order to eliminate unknown phase behaviour in both loudspeaker and microphone. A limitation which unfortunately could not be quantified in the test flight described here is probably imposed by sound carried by the frame once the loudspeaker efficiency is reduced at low pressures. This effect was countered by compensation microphones sensitive only to signals in the structure. The received signals were restored in phase locked loops (PLL) and the phase between the microphones established digitally by gating with 100 MHz. The on-board processor formed the mean of the propagation times t . . . . = (tl + tz ) / 2 over the distances in opposite directions with the intention to (largely) eliminate errors due to Doppler shift caused by possible cross winds. The resulting velocity Vc~tcutated = d / t .... is to a very

.I1+ - -l / . Furthermore these results are averaged to a basic data

good approximation the mean velocity v .... = - - - 2 ( t,

t2 )

rate of about 100 s -~ to be stored on board. Algorithms are employed to reject outliers and to assure that propagation time variations which exceed 360 ° in phase are averaged correctly. For a detailed description of the instrument see Posch (1998) and Posch et al. (1999). TEST FLIGHT The instrument was built as an autonomous unit with its own data-logger and a conventional (non-meteorological) thermistor was added to the package. The balloon was launched on September 6, 1999, from ESRANGE, Sweden, at 07.11 UT. 30 Minutes later a commercial ozone sonde which also carried a meteorological temperature sensor was flown. Unfortunately the MST radar at ESRANGE had been hit by lightening and performed only very poorly. Figure 1 shows the meagre echoes received on that day vs. time together with the balloon trajectories of the acoustic and ozone payloads, respectively. The balloon of the acoustic payload under-performed and only reached 18 an apogee of 19 km, whereas the ozone sonde reached the full 16 anticipated altitude in excess of 30 km. Comparisons of the acoustic 14 sensor up to an altitude of 19 km are confined to (a) the data of an ~12 inappropriate thermistor on the same payload and (b) to data from a -~10 proper meteorological sensor flown < 30 minutes later. Using the nominal 8 geometry of the acoustic instrument (i.e. 50 cm path) yielded temperatures about 4 K higher than the ones obtained from the on-board thermistor, which is mounted inside a ventilated sunshade on the side of the payload about 1 m from the path of the acoustic sonde. Therefore the 6 7 8 9 10 11 12 real temperature difference between UT the two locations is small compared Fig. 1. Echoes of the MST radar (crippled by lightning) as a function of to the measured difference. In the time around the balloon flights. White line: acoustic payload, black line: first few kilometres it is assumed upleg of the ozone sonde. that the measurement with the

High Resolution Temperature Sensor 20

1367

i

~ r

~--

_% ~

12

.

.

.

.

.

.

.

"

-

.

/

i

.

acoustic (mean 10 s) [

/

~ . . . . . . . . . . . .

r

. . . . . . . . . . . .

! ............

L ....

,. . . . . . . . . . . .

r

. . . . . . . . . . . .

,. . . . . . . . . . . .

o=one soodo

/

T . . . . . . . . . . . .

~. . . . . . . . . . . .

i ............

::............

E "O

10

:3

<

...........

6

......

L ............

............

!!!!!!!!!!!! !!!!!

!

2

0 210

I

I

I

I

I

220

230

240

250

260

270

280

290

Temperature, K Fig. 2. Temperature profiles of the acoustic instrument and of the ozone sonde 30 minutes later. The data of the acoustic sonde are 10 s averages in order to make the resolution comparable to that of the commercial ozone sonde.

thermistor is trustworthy because of sufficient thermal contact due to the dense atmosphere. The effective acoustic length needed for agreement with results of the on-board thermistor was only about 5 mm less than the nominal. All further processing is done with the reduced (effective) distance. At higher altitudes the thermistor showed values significantly larger (10 K) than both the data from the acoustic instrument and from the ozone sonde flown half an hour later. We therefore conclude that we must disregard the thermistor as not suitable above say 3 km. The descent from the apogee at 19 km by parachute was faster than the ascent and - more importantly - probably less stable. The downleg data have more erratic values, but the mean behaviour shows a temperature profile in agreement with the upleg data; we will therefore restrict the analysis to the upleg. Figure 2 shows the temperatures from the acoustic instrument (using the corrected path length) and the data from the ozone sonde. The acoustic data are 10 s averages to yield a resolution comparable to the conventional radiosonde. The general agreement up to 11 km is satisfactory, the slightly different heights of the three temperature inversions in the troposphere can very well be due to either temporal or spatial variations. In Figure 3 we form the ratio between the values of the two instruments. Significant departures only occur above 12 km and again could very well be real. Having gained confidence in the data we will now look into the high resolution plots. In Figure 4 the inversion near 4.5 km is shown in successive degrees of resolution. The full data (without averaging) is one measurement each 5 cm of altitude. However also the 0.1 s (50 cm) values show large gradients, but also suspicious periodicities. Similarly large gradients have been observed by Dalaudier et al. (1994) on a number of flights from southern France. The significance of the high resolution data is checked by performing a (de-trended) spectral analysis shown in Figure 5 taken over the whole ascent. The slope is a little shallower than the -5/3 gradient expected from turbulence theory and there is peak at 20 Hz, a feature found at all altitudes.

1368

M. Friedrich et al.

2o

! i

18 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

i ..........

~

'° . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i ~ - i

14

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-' . . . . . . . . . . . . . . .

.

-~

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.....5 . . . . . . . . . . . . . . . . . . . . . . .

- ;- -- - ~

12 .................................. ~

.

........

................................................

E "0

10................... ~ - i

.................................................

i ................

<

4 .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

' .

.

.

.

.

.

~ ......................... 0 0.98

.

.

.

.

~

~

-_____E---:. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

~

..........................

i

i

I

8.99

I

101

~

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

i ............................... ~ i 102

I03

1 04

Temperature ratio (conventional/acoustic) Fig. 3. Ratio between the acoustically derived temperatures to the ones from the instrument aboard the ozone sonde.

CONCLUSIONS The flight demonstrated that temperatures can indeed be derived with very high resolution at least up to an altitude of 19 km; whether body-borne sound really poses problem at higher altitudes remains to be tested. Improvements are feasible in the electronics such as to control the necessary acoustic power in a loop (power saving), more importantly the PLL restoring the signals from the microphones could be replaced by a direct correlation between the (analogue) signals. This should eliminate the 20 Hz in the spectrum which we attribute to the frequency of the low pass in the PLL. REFERENCES Dalaudier, F., C. Sidi, M. Crochet, and J. Vernin: Direct Evidence of "Sheets" in the Atmospheric Temperature Field, J. atmos. Sci. 51 (2), 237-248, 1994. Luce, H., M. Crochet, F. Dalaudier, and C. Sidi: Interpretation of VHF ST Radar Vertical Echoes from in sittt Temperature Sheet Observations, Radio Sci. 30 (4), 1,000-1,025, 1995. Posch, M.: Messung der Schailgeschwindigkeit zur tragheitslosen Bestimmung der Lufttemperatur, MSc Thesis, Technical University Graz, 1998. Posch, M., M. Friedrich, S. Kirkwood, and K. Stebel: An Acoustic Balloon-Borne Instrument to Measure Temperature Fine Structure Near the Tropopause, ESA SP-437, 359-362, 1999.

High ResolutionTemperatureSensor 51 , ~ - -

!

..!...............]__

1369 !

! acoustic acoustic (mean 1 s) acoustic (mean 113 s) ozone sonde

_

_

49

4B

d w,.=

47

< 4.6

45

44 ~ 263

2635

2S4

264.5

265

2(~5.5

266

2665

267

267 5

2~8

Temperature, K Fig. 4. Inversion layer near 4.5 km with successively higher resolution (0.01, 1, 10 s-I). 10"

106

b 0 >

10 .3

I0 "z

10 "I

10 o

101

Frequency, Hz

Fig. 5. Temperature spectrum derived from the whole ascent (de-trended). Note the peak near 20 Hz which is probably an instrumental effect from the PLL. The -5/3 slope appears to exist up to frequencies of 0.06 Hz (or an altitude scale of about 100 m). The steeper spectrum above 0.1 Hz is from lower altitudes (<11 km), the shallower one from above.