Periodic Reporting for period 1 - AVATAR (Transformative Digital Air Vehicle with IoT Sensors for Safer Urban Skies)
Reporting period: 2023-02-01 to 2024-07-31
AVATAR’s objective is to create a Digital Twin (DT) to continuously monitor the operation of transformative air vehicles (TAV) through an innovative smart IoT sensing skin, informed by accurate physical models and advanced analytics. AVATAR’s DT framework builds on verified physical models for TAVs design, with the potential to contribute to an extended safe-life design of TAVs.
The 4 key objectives include:
1. IoT Sensing Skin: Development of a multifunctional, self-powered smart skin embedded with sensors that provide real-time data on structural health (strain, acceleration, temperature, etc.).
2. Digital Twin: Creating a virtual model that mirrors the physical air vehicle, predicting its behaviour, and providing early warnings of potential failures.
3. Advanced Analytics: Implementation of machine learning (ML) and artificial intelligence (AI) techniques for fault detection, remaining useful life (RUL) estimation, and flight profile optimization.
4. Communication and Cloud Platform: Ensuring secure data transmission from the vehicle to a cloud platform for real-time processing and storage, while maintaining high levels of data security and reliability.
The expected impacts of AVATAR include: i) reduced operational costs through predictive maintenance, ii) improve safety of urban air mobility, and iii) providing insights to optimize vehicle design for future generations.
The 4 key objectives include:
1. IoT Sensing Skin: Development of a multifunctional, self-powered smart skin embedded with sensors that provide real-time data on structural health (strain, acceleration, temperature, etc.).
2. Digital Twin: Creating a virtual model that mirrors the physical air vehicle, predicting its behaviour, and providing early warnings of potential failures.
3. Advanced Analytics: Implementation of machine learning (ML) and artificial intelligence (AI) techniques for fault detection, remaining useful life (RUL) estimation, and flight profile optimization.
4. Communication and Cloud Platform: Ensuring secure data transmission from the vehicle to a cloud platform for real-time processing and storage, while maintaining high levels of data security and reliability.
The expected impacts of AVATAR include: i) reduced operational costs through predictive maintenance, ii) improve safety of urban air mobility, and iii) providing insights to optimize vehicle design for future generations.
To reach the above 4 main objectives, AVATAR is structured around a building block (BB) approach, ensuring a systematic progression from concept to deployment. This approach includes two parallel tracks: 1) the data collection from experimental setups and field activities and 2) the development of an accurate digital twin of the aircraft's structure. Each track has three levels, each building upon the previous.
The progress of work to date towards each objective is summarised below against the expected technology readiness levels (TRLs), and more in detail in section 1.2 of Part B:
1. IoT Sensing Skin: The IoT sensing skin has two main developments that run in parallel: the multi-functional sensors (including wiring, installation and connectors) and a miniaturised data acquisition (DAQ) unit that includes wireless communication and energy harvesting modules. Both developments follow a building block approach to reach TRL5. The multi-functional sensors include novel printed strain sensors and thermistors, accelerometers and Piezoelectric transducers. These sensors have been tested at level 1 demonstrating TRL 2-3 for their functional requirements. They are currently under test at level 2, to demonstrate TRL 4 (see Deliverable 2.1 for more information). The sensors that have been selected and tested for multi-parameter sensing include: printed strain sensors and thermistors, and commercial accelerometer and PZT transducers. The developed DAQ includes a microcontroller unit (MCU) that records and processes data directly coming from the sensors of the sensing skin, tested following the plan in Table 1 of Part B description.
2. Digital Twin: Currently there are three parallel activities that are ongoing that will provide feedback to AVATAR’s DT implementation. The first activity involves the structural health monitoring of the structure. The algorithms implemented have been demonstrated in representative coupons and have reached TRL3. Now, the methodologies are tested on a wing from ASTERO drone and it is planned to also test them on EVE’s wing spar component test in Q1 2025 to achieve TRL 4. The second activity involves load monitoring based on on-line data from sensors. Based on the proposed approach, the stress field is successfully reconstructed based on measurements from a limited number of strain gauges for both static and dynamic test cases, reaching a TRL 3. Finally, the third activity under progress is the implementation of machine learning for flight phase identification. The method is applied directly to data from actual flight tests, and it has reached TRL 2 – 3. Over the next reporting period, the methods will be progressed to the higher levels of the building block as well as validating and demonstrating their performance.
3. Advanced Analytics: The development of advanced analytics follows closely the digital twin developments. This includes the development of models that support the computational requirements of the digital twin modules. In this reporting period, the following algorithms have been developed:
a. Impact detection and force reconstruction for passive sensing
b. Self-organizing maps for flight phase characterization
c. Operator learning neural networks for surrogate modelling of simple aluminium coupons
All algorithms have been tested using simulated data and therefore are expected to have achieved TRL2. Over the next reporting period the algorithms will be tested against measurements collected from the component tests and demonstrate TRL 3-4.
4. Communication and Cloud Platform: For the communication subsystem, an initial data transfer between the sensor node and the gateway has been developed and tested in the lab. On the other hand, discussions around integration of the communication subsystem with DAQ in the sensor node and the communication with cloud data transfer module in the gateway is ongoing.
The progress of work to date towards each objective is summarised below against the expected technology readiness levels (TRLs), and more in detail in section 1.2 of Part B:
1. IoT Sensing Skin: The IoT sensing skin has two main developments that run in parallel: the multi-functional sensors (including wiring, installation and connectors) and a miniaturised data acquisition (DAQ) unit that includes wireless communication and energy harvesting modules. Both developments follow a building block approach to reach TRL5. The multi-functional sensors include novel printed strain sensors and thermistors, accelerometers and Piezoelectric transducers. These sensors have been tested at level 1 demonstrating TRL 2-3 for their functional requirements. They are currently under test at level 2, to demonstrate TRL 4 (see Deliverable 2.1 for more information). The sensors that have been selected and tested for multi-parameter sensing include: printed strain sensors and thermistors, and commercial accelerometer and PZT transducers. The developed DAQ includes a microcontroller unit (MCU) that records and processes data directly coming from the sensors of the sensing skin, tested following the plan in Table 1 of Part B description.
2. Digital Twin: Currently there are three parallel activities that are ongoing that will provide feedback to AVATAR’s DT implementation. The first activity involves the structural health monitoring of the structure. The algorithms implemented have been demonstrated in representative coupons and have reached TRL3. Now, the methodologies are tested on a wing from ASTERO drone and it is planned to also test them on EVE’s wing spar component test in Q1 2025 to achieve TRL 4. The second activity involves load monitoring based on on-line data from sensors. Based on the proposed approach, the stress field is successfully reconstructed based on measurements from a limited number of strain gauges for both static and dynamic test cases, reaching a TRL 3. Finally, the third activity under progress is the implementation of machine learning for flight phase identification. The method is applied directly to data from actual flight tests, and it has reached TRL 2 – 3. Over the next reporting period, the methods will be progressed to the higher levels of the building block as well as validating and demonstrating their performance.
3. Advanced Analytics: The development of advanced analytics follows closely the digital twin developments. This includes the development of models that support the computational requirements of the digital twin modules. In this reporting period, the following algorithms have been developed:
a. Impact detection and force reconstruction for passive sensing
b. Self-organizing maps for flight phase characterization
c. Operator learning neural networks for surrogate modelling of simple aluminium coupons
All algorithms have been tested using simulated data and therefore are expected to have achieved TRL2. Over the next reporting period the algorithms will be tested against measurements collected from the component tests and demonstrate TRL 3-4.
4. Communication and Cloud Platform: For the communication subsystem, an initial data transfer between the sensor node and the gateway has been developed and tested in the lab. On the other hand, discussions around integration of the communication subsystem with DAQ in the sensor node and the communication with cloud data transfer module in the gateway is ongoing.
The AVATAR IoT sensing skin is a lightweight, energy-efficient, modular technology offering a flexible multi-sensing solution for health monitoring and predictive maintenance. Current industrial data acquisition systems are bulky, costly, and lack integration for multiple sensors like PZT, strain gauges, thermal sensors, and accelerometers. AVATAR bridges this gap with its novel flexible sensor system and tailored data acquisition solution, compatible with advanced cloud-based platforms like AWS for secure transmission, processing, and storage. Compliance with European data strategies, ensured by NTT and UL, positions AVATAR for industrialization as a potential new product.
Building on these advances, AVATAR’s Digital Twin (DT) platform pushes predictive maintenance and structural health monitoring for air vehicles, providing two key scientific contributions:
1. Flexible and Configurable Solution: Adaptable to different structures, AVATAR supports varied sensor configurations for composite and metallic materials, enhancing versatility.
2. Multilevel Monitoring Strategy: Integrating in-flight monitoring with on-ground assessments, AVATAR delivers accurate, comprehensive structural health evaluations, boosting safety and reliability.
These contributions are validated through AVATAR’s building block approach, progressing in complexity. The final demonstrators—from a small, manned aircraft to a light UAV—test the solution in diverse settings. The small aircraft needs extensive sensing coverage, while the UAV has weight, size, and composite structure constraints.
Building on these advances, AVATAR’s Digital Twin (DT) platform pushes predictive maintenance and structural health monitoring for air vehicles, providing two key scientific contributions:
1. Flexible and Configurable Solution: Adaptable to different structures, AVATAR supports varied sensor configurations for composite and metallic materials, enhancing versatility.
2. Multilevel Monitoring Strategy: Integrating in-flight monitoring with on-ground assessments, AVATAR delivers accurate, comprehensive structural health evaluations, boosting safety and reliability.
These contributions are validated through AVATAR’s building block approach, progressing in complexity. The final demonstrators—from a small, manned aircraft to a light UAV—test the solution in diverse settings. The small aircraft needs extensive sensing coverage, while the UAV has weight, size, and composite structure constraints.