Multi-sensor and multi-mode technologies for optimized data acquisition

CEA-List is developing multi-sensor and multi-mode technologies for the analysis of large manufactured parts and continuous measurement—new capabilities that will expand what is possible in the testing, inspection, and control of structures, parts, and processes.

CEA-List is leveraging two key technologies: multi-sensor, where multiple sensors are installed across a structure, for example, and multi-mode, where multiple industrial process parameters are measured.

When used in combination with imaging, data fusion, AI, and other algorithms, these technologies will enable new part inspection, structural health monitoring, process control, and environmental monitoring solutions across a wide range of industries. They are of particular interest in the field of non-destructive testing (NDT).

Key use case: Structural Health Monitoring (SHM)

Structural Health Monitoring (SHM) is a means of detecting early-stage damage in complex structures like buildings, road and rail infrastructure, and aircraft. SHM systems use arrays of permanently installed sensors to measure various structural parameters, either continuously or at defined intervals. The measurements are then analyzed to detect and monitor changes in the structure’s geometric and material properties. SHM has the potential to replace periodic manual inspections by human operators, reducing maintenance costs and downtime and increasing safety.

CEA-List’s innovative SHM solutions use guided elastic waves that can propagate over long distances, making them ideal for monitoring large structures such as buildings and bridges. The arrays of sensors transmit signals, which are processed using algorithms and AI techniques to detect damage or generate images of areas requiring inspection. The institute is also working on passive guided-wave monitoring techniques for more challenging environments, using optical fibers as ultrasonic wave receivers to detect acoustic emissions from cracks and other defects.

CEA-List is home to an open research and innovation platform for SHM technologies (SACHEMS), where new SHM solutions are developed and tested.

SHM for the aerospace industry, in partnership with Safran

CEA-List has partnered with Safran to develop an IA-based SHM system to monitor defects in composite aircraft structures. The system uses a neural network that has been trained to detect early signs of damage by interpreting guided-wave images. The teams used simulated data rather than experimental data to train the network and create a rapid, reliable diagnostic tool.

Find out more about Structural Health Monitoring in episode 4 of the Expert Insights podcast (in French), where we talk to Research Director and NDT expert Pierre Calmon.

Proof-of-concept prototypes and demonstrators

Our researchers are also building proof-of-concept prototypes for multi-mode technologies that leverage data fusion algorithms. Early applications will include industrial process control and nuclear instrumentation. Research into the next generation of miniaturized low-power measurement sensors for use in extreme environments is also underway.

Our technologies

CEA-List uses a wide range of different technologies, for example:

  • Ultrasound
  • Guided waves
  • Eddy currents
  • X-ray
  • Gammagraphy
  • Tomography
  • Visible and infrared cameras

European programs

We leverage our deep knowledge of use cases, digital models, and algorithms to optimize the design, implementation, and operation of sensor solutions, whatever technologies we’re using and whatever kind of tool or structure—like a robot, for example—we’re equipping. We also strive to make the solutions as non-intrusive as possible.

Pierre Calmon

Research Director — CEA-List

See also

Research programs

Optimized, interoperable models

Simulation is making inroads into all aspects of industrial processes, making optimized, interoperable models a key issue.
Read more
Research programs

Diagnostics and prognostics

Measurement, data analysis, simulation, and machine learning are vital to understanding industrial processes, predicting their performance, and automating decision making.
Read more