Development of an in situ Acoustic Anemometer to Measure Wind in the Stratosphere for SENSOR

The Stratospheric Environmental respoNses to Solar stORms (SENSOR) campaign 10 investigates the influence of solar storms on the stratosphere. This campaign employs a long-duration zero-pressure balloon as a platform to carry multiple types of payloads during a series of flight experiments in the mid-latitude stratosphere from 2019 to 2022. This article describes the development and testing of an acoustic anemometer for obtaining in situ wind measurements along the balloon trajectory. Developing this anemometer was necessary, as there is no existing commercial off-the-shelf 15 product, to the authors’ knowledge, capable of obtaining in situ wind measurements on a high-altitude balloon or other similar floating platform in the stratosphere. The anemometer is also equipped with temperature, pressure, and humidity sensors from a Temperature-Pressure-Humidity measurement module, inherited from a radiosonde developed for sounding balloons. The acoustic anemometer and other sensors were used in a flight experiment of the SENSOR campaign that took place in the Da chaidan 20 District (95.37°E, 37.74°N) on 4 September 2019. The zonal and meridional wind speed observations, which were obtained during level flight at an altitude exceeding 20 km, are presented. This is the first time that in situ wind measurements were obtained during level flight at this altitude. In addition to wind speed measurements, temperature, pressure, and relative humidity measurements during ascent are compared to observations from a nearby radiosonde launched four hours earlier. Further analysis of the 25 wind data will presented in a subsequent publication. The problems experienced by the acoustic anemometer during the 2019 experiment show that the acoustic anemometer must be improved for future experiments in the SENSOR campaign.

For the SENSOR campaign, we have developed an acoustic anemometer that can be used with a long- analyses of the measurements yielded by the anemometer will be discussed in subsequent publications.
This article is arranged as follows. The principle of operation of acoustic anemometers and the development of this acoustic anemometer are introduced in Section 2. A detailed description of the 2019 85 balloon-borne experiments and preliminary results are presented in Section 3. Conclusions regarding the acoustic anemometer are described in Section 4.

Principle of operation
An acoustic anemometer is used to measure wind speed by sensing the difference in propagation time of 90 the sonic signal in the windward and leeward directions caused by the movement of airflow (Coppin and Taylor, 1983;Alberigi Quaranta et al., 1985;Fernandes et al., 2017). Taking the measurement of onedimensional wind velocity, for example, there is a pair of transducers that are facing to each other with a distance of L (as shown in Fig. 1). Each transducer can function as a transmitter as well as a receiver.
To measure wind speed, each transducer transmits acoustic signals, and the opposite transducer is used 95 as a detector to receive the signals. Due to the airflow, the flight time of sound waves in opposite directions between the pair of transducers, denoted as and , differs. The values of and have the following relationships with the wind velocity in the along-transducer direction (denoted as ): Here, represents the travel time of signals in the windward (against the wind) direction, while represents the travel time in the leeward (with the wind) direction. C is the speed of sound. Because and can be directly measured, can be derived as follows: It should be mentioned that when the anemometer is on a high-altitude balloon, it measures the wind speed relative to the motion of the gondola rather than the earth-relative (absolute) wind speed. Thus, the absolute wind speed, denoted as , is the sum of the speed obtained by the anemometer and the speed of the gondola's motion ( ). If the relative wind is measured in the same direction as the gondola's motion, then can be expressed as:

Instrument Design
The acoustic anemometer is one of the payloads carried by the high-altitude balloon during the SENSOR campaign. It is mainly comprised of two parts: the sensors mounted outside the balloon gondola through an aluminium alloy boom, and the electronics box installed inside the gondola. Each part is described in 115 detail below.
The acoustic anemometer employs three pairs of ultrasonic transducers arranged in a three-dimensional structure to measure wind. The balloon flies in an environment of low temperature and low pressure, causing the acoustic signals to experience more attenuation than they would in the lower troposphere. This attenuation increases with increasing frequency (Sutherland and Bass, 2004). To the authors' 120 knowledge, most commercial acoustic anemometers operate at frequencies above 100 kHz, which would result in severe attenuation of acoustic signals under the conditions experienced in the stratosphere during SENSOR, possibly even making the signals indistinguishable from background noise. To improve the signal-to-noise ratio (SNR) as much as possible with a distance between transducers of 0.2 m, we have chosen ultrasonic transducers that operate at a lower frequency of 40 kHz. This is the primary difference 125 between the anemometer that we developed and the anemometers used in previous high-altitude balloon studies referenced in the Introduction. The received signal is amplified immediately by an ultra-low noise preamplifier to further improve the SNR, and an Automatic Gain Control (AGC) circuit is also used, different from terrestrial anemometers, to adjust its gain levels by altitude range, because the received signal decreases as altitude increases. The different gain levels are determined by ground testing in a 130 vacuum chamber.
The transducers are installed on a bracket with ring structures, which are manufactured by 3-D printing to ensure that each transducer is aligned with the opposite one to maximize the SNR and to ensure that the distance between transducers are precise. The preamplifier and AGC circuits are located at the bottom of the bracket instead of the electronic box to avoid transmission loss of the signal caused by the long 135 cable between transducers and the electronic box.
During flight, temperatures outside of the gondola can drop to as low as −70°C , with the lowest temperatures occurring when the balloon passes through the tropopause. Therefore, the transducers and electric devices used in the preamplifier and AGC circuits were chosen for a wide temperature range and were tested in a thermal vacuum chamber at temperatures as low as −70°C to ensure that the transducers 140 and circuits function under such an environment. This extreme environment has lower temperatures than the stratosphere in which the anemometer is used in SENSOR. To further protect against extreme conditions, the spaces where the circuits are placed were also been insulated.
The amplified signals, connected to the electronic box inside of the gondola by long cables, are generated at 1 MHz frequency by the Analog-to-Digital Converter (ADC) on a controller unit, which is one of the 145 three boards in the electronic box. The controller board serves as a "brain" in the electrical system of the anemometer. Its core is an onboard FPGA, which operates the normal workflow of the anemometer. It generates a pulse train for the transmitting transducer, which is outputted to a Digital-to-Analog Converter (DAC), and then amplified to about 90 V (peak to peak) by the relevant driver circuits, the second board in the electronic box. The controller board also adjusts the gain levels of the AGC circuits 150 through a DAC according to the gondola's current altitude. Finally, the controller board employs the communication interface, RS422, with gondola's storage system; however, due to limited bandwidth, only the health data used to monitor the anemometer and a portion of the observation data can be delivered to the gondola's storage system. As the anemometer is recovered after each flight, we store the observation data sampled by the ADC on a large-capacity storage card. This card also contains other 155 datasets, such as the command data received by the RS422 interface from gondola's flight computer, which includes GPS time, altitude, longitude, latitude, gondola attitude, and gondola speed data.
The third board is the power supply, which converts unregulated +28 V power provided by the gondola to regulated +12 V and +8 V power for use by the other circuits. A fuse in parallel with another fuse and a power resister are added to the input to protect the +28V power supply in case the input current exceeds 160 5A. The DC/DC converter is protected by a surge protection circuit that limits the inrush current and the start-up voltage slew. An additional electromagnetic interference (EMI) filter is also used on the power lines. In addition to wind speed, the anemometer incorporates sensors to measure temperature, pressure, and humidity, which are located on the bracket outside of the gondola. The measurements are delivered to 165 the controller board through an RS232 interface on a Temperature-Pressure-Humidity measurement module (TPH module), which was inherited from a radiosonde that we developed and used in a sounding balloon. We retained the same circuits and sensors here. The temperature, pressure, and relative humidity data were also stored on the large-capacity storage card.

Data Processing 170
In the current experimental design, we store the raw sampled data and perform post-processing when the anemometer is recovered, rather than conducting real-time online processing. To obtain wind speed following the measurement method described in Section 2.1, the acoustic signal propagation time between transducers should be determined first. However, due to electronic delay, the acoustic signal propagation time is not measured directly as the data is obtained. Instead, the ultrasonic waves received 175 by each transducer when there is no wind are obtained first and are stored in the storage card as the reference signals. When the anemometer experiences wind, there are time differences between the received signals and the reference signals, allowing calculation of wind speed without knowledge of the exact electronic delay. The mathematical model employed can be illustrated as follows.
When the wind speed is zero, the measured travel time of ultrasonic waves between one pair of 180 transducers can be inferred as where 0 is the speed of sound in the current environment, ℎ is the electronic delay along the transmitted signal path, and 0 is the time of flight (TOF) without electronic delay.
When the transducers are subjected to wind, the measured signal travel time in the leeward direction is 185 where 1 is the speed of sound in the current environment, v is the wind speed along the pair of transducers, and 1 is the TOF without electronic delay in the leeward direction. The time difference between the received signal and the reference signal can be obtained by adopting cross correlation between the received ultrasonic wave and the corresponding reference wave: 190 Thus, 1 would be: Therefore, from Eqs. (4-7) TOF can be obtained without knowing the exact electronic delay. Similarly, 2 , which is the TOF in the windward direction, can be obtained from: 195 The wind speed can then be obtained by substituting Eqs. (7) and (8)