Instrumentation & Measurement Magazine 26-4 - 15

However, they do not directly capture the aerodynamic performance
essential for aeroelastic instabilities, whereas local
measurements on the blade measuring the aerodynamic forces
can further enable tracking of deterioration and furnish diagnostics
to prevent failures and long repairs. Monitoring
through novel sensing solutions can facilitate decisions to support
shorter downtime and optimized maintenance schemes,
which eventually lower the cost of energy, especially for wind
turbines installed in remote and offshore locations. Moreover,
extracting information from several wind turbines within a
farm can support the design and operation to increase efficiency
at the farm level.
The extraction of aerodynamic data is further informative
on the condition of blade components. The state-of-the-art
aerodynamic measurement systems installed on rotor blades
in real conditions employ high-quality and costly sensors,
with pressure taps and long tubes, and/or optical fibers running
inside the blade [9]. These measurement campaigns have
demonstrated the potential value of local aerodynamic measurements
on the rotor blade but are limited by the complexity
and the cost of such embedded sensor systems, which are not
viable for cost-effective prototyping and monitoring system
in the industry. If aerodynamic measurements are to be useful
and viable in the industry, they can only be accomplished via
robust and cost-effective measurement systems that are easily
installed on the blades.
Aerosense System Architecture and
Low-power SCADA Sensor Node
In this work we focus on a sensing system for addressing the
needs for cost-effective wind turbine monitoring and data
analysis solutions, which can offer insights into the complex
aerodynamic and acoustic behavior of large, flexible blades.
The incoming flow transports turbulent structures of different
scales spatially and temporally, resulting in complex aerodynamic
load fluctuations that are difficult to simulate and
therefore require validation in the field.
The Aerosense system (Fig. 1) is designed to collect aerodynamic
and acoustic data directly on operating wind
turbine blades. The system includes three main components:
a versatile, self-sustaining, and scalable sensor node;
a local BLE gateway connected with the internet; and a remote
digital twin application running on a cloud server.
These three main blocks are specifically designed to fulfil
the Aerosense project scope, enabling high precision
measurements and automated data analysis and system
modeling using advanced machine learning techniques,
such as graph neural networks (GNN). Moreover, all components
are optimized to be cost-effective, ultra-low power
and robust, to enable long-term and continuous monitoring
in harsh environments, such as in the presence of ice or
strong solar irradiation.
June 2023
The digital twin is a crucial component of the project, as it
enables the utilization of measurement data to provide added
value to the system users. As depicted in Fig. 1, the digital
twin encompasses the following elements: a data pre-processor
responsible for collecting, timestamping, cleaning,
correcting, calibrating and storing measurement data as well
as external information (such as SCADA acquisition and metadata)
in a database; inverse problem solvers utilized to infer
quantities such as the angle of attack, full flow field reconstruction,
and leading edge erosion class, through the use of
trained machine learning models; and forward problem solvers
employed to predict non-measured quantities such as
structural deformation of the blade through the application
of fluid-structure-interaction simulations; data analysis algorithms,
including a post-processor, which computes derived
quantities such as lift and drag coefficients; and dashboards
employed to display and provide access to the results with an
user-friendly interface.
To gather data from low-power devices, working as a local
computing support and as a protocol bridge, a custom
base station was designed. It supports the local BLE wireless
link and the time synchronization among devices and
provides a local storage buffer in case of internet connection
drops. Moreover, the base station is the local coordinator of the
measurement system, and triggers the sensor node working
modalities based on a device-specific configuration file that is
automatically downloaded from the server (digital twin) once
a stable internet connection is available. The acquisition interval
can be programmed, while it is further possible to decide
on the set of devices and specific sensors to be activated and
modified in runtime, between measurement sequences. As
shown in Fig. 1, three sensors installed on a single blade can
be synchronously activated, with a timestamp accuracy down
to 100 μs, independently from other turbines, allowing for
user-activation upon request for the entirety or a subset of installed
sensors; for example, barometers, differential pressure
sensors, and microphones can be controlled and activated/deactivated
individually on each sensor node. On the other side,
a single sensor node can be mounted on a wind turbine, or considering
another possible installation setting, the Aerosense
system can cover all the blades of a turbine tower.
The sensor node hardware schematic is presented in
Fig. 2. It comprises five blocks, one main rigid board (gray
background), three flexible 0.2 mm thin printed circuit boards
(PCB) to support a heterogeneous set of MEMS sensors, and a
flexible energy harvesting circuit composed by a solar panel.
The core of the system relies on a System on Chip (SoC) from
Texas Instruments including three microprocessors, one ARM
Cortex M0 fully dedicated to the BLE protocol, one 8-bit core
to support digital sensors, and a Cortex M4 as main coordinator
and to support on board processing. The CC2553P SoC is
coupled with a non-volatile memory of 512 MB, used as a data
IEEE Instrumentation & Measurement Magazine
15
Instrumentation & Measurement Magazine 26-4

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