Instrumentation & Measurement Magazine 26-4 - 17

intrinsic redundancy in
the transmitted data. This
peculiarity opens several
possibilities for future research
directions, such as
onboard data compression
to decrease the wireless
transmission energy or the
possibility of interpolating
measurements in case of
unstable or malfunctioning
barometers.
Conclusion
Fig. 3. Real installation during winter. (a) The Aerosense prototype was installed on the blade of a 6kW wind turbine.
(b) The sensor node was installed on the blade. The bottom sensor strip comprises an array of 40 barometers, the top one
contains the main sensor with the flexible solar panel and the battery, and the middle section encapsulates five differential
pressure sensors. (c) The sensor node was decorated with white lines to ease the photogrammetry process and retrieve the
3D shape of the section and positions of the sensors.
Fig. 4. Time resolved measurements extracted from the BigQuery database. (a) Main direction of the accelerometer
indicating the rotational speed of the wind turbine. (b) 40 pressure measurements around the blade, used to compute the
aerodynamic force.
force, thus also correlated with the rotational speed. Fig. 4 also
shows pressure data (raw format) collected during the same
period with an array of 40 barometers. Here, the visible oscillatory
contribution is given by the rapid change in blade
altitude, while the average value is a contribution of the ambient
air pressure and the wind flow tangentially to the blade.
Extracting the latter from the raw data provides the possibility
to infer the pressure distribution along the blade surface,
thus generating precise models in real and diverse operational
conditions. Notably, the high correlation among barometers
is easily visible due to their spatial proximity, generating an
June 2023
This work overviews the
Aerosense sensing system
for wireless, non-intrusive,
low-cost aerodynamic and
acoustic measurements
on operating wind turbines.
It demonstrates the
possibility of measuring
operational parameters directly
from a blade surface
without needing external
infrastructure or costly custom
installations. This work
overviews the requirements
for the acquisition
of quantities defined for
the Aerosense downstream
tasks, namely the characterization
of the blades at
the individual turbine level
and, eventually, the creation
of real-time tracking
digital twins, which can
be exploited for decision
support at the wind farm
level. Future system development
will focus on
delivering more intelligence to the sensor node, specifically regarding
onboard lossy data compression algorithms and early
detection models to decrease the system response latency.
Moreover, transmitting only relevant features to the digital
twin server supports system scalability, where hundreds of
sensor nodes can eventually be deployed on a wind farm.
Acknowledgment
This work was supported in part by the Swiss National Science
Foundation (SNSF) Bridge Project " AeroSense " under
Project 40B2-0_187087. Moreover, authors sincerely thank
IEEE Instrumentation & Measurement Magazine
17

Instrumentation & Measurement Magazine 26-4

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 26-4