An autonomous four-legged robot serving industrial maintenance: Léo Chastain’s final-year project at Safran Blades

The aerospace industry has particularly high requirements in terms of quality, safety, equipment availability and process control. It was within this context that Léo Chastain completed his final-year project at Safran Blades, a company specialised in the production of aerospace components.

Based in Marchin on a modernised former industrial site, Safran Blades is undergoing an industrial transformation in which automation, robotics and data are playing an increasingly important role in production and maintenance.

Léo's project had a very concrete objective: to transform Spot, the robot developed by Boston Dynamics, into an inspection tool that could be effectively deployed in an industrial environment.

Spot is much more than a demonstration robot. It is an autonomous mobile platform capable of navigating through a factory, climbing stairs, following predefined routes and performing various analyses of industrial equipment. During the project, the robot was configured to carry out inspection rounds for maintenance, safety and security, based on thermal, acoustic and visual analyses.

Léo worked on several aspects of the project. First, he had to gain an in-depth understanding of the robot's technical ecosystem: the Orbit platform, the control tablet, the integrated equipment, sensors and fiducials used for localisation, as well as the capabilities and limitations of the associated software.

This learning phase was necessary to create, adapt and optimise Spot's missions in a real industrial environment.

A significant part of the work then focused on defining inspection routes and the analyses to be performed along them. On site, Léo identified equipment and components that could potentially show anomalies, such as motors, pumps, compressed-air networks, pressure gauges, busbars, air heaters, fire extinguishers and hose reels.

For each of these components, it had to be determined exactly what Spot should observe, measure or check. The robot was programmed, among other things, to detect overheating, air leaks, acoustic anomalies and abnormal readings on meters and pressure gauges.

The project illustrates an important evolution in industrial maintenance. It is no longer only about intervening when equipment fails, but also about collecting frequent and relevant observations in order to detect anomalies as early as possible.

Spot makes it possible to carry out inspections more frequently, repeatedly and with greater traceability. The robot does not replace the expertise of maintenance technicians. Instead, it provides them with more regular and structured information that can be more easily integrated into a maintenance process.

An important part of the final-year project was therefore the structuring of the data collected. In order for an anomaly detected by the robot to be genuinely useful, it needs to be linked to the correct asset.

Léo therefore worked on assigning Asset IDs: unique identifiers that allow Spot's observations to be linked to the corresponding components in maintenance systems. This means that an air leak, acoustic anomaly or abnormal temperature can be linked to a specific asset and eventually transferred to the CMMS.

Integration with the CMMS was an important part of the project. During the internship, fully automated transfer of anomalies to Coswin was not yet available. Léo therefore manually registered work orders so that the anomalies detected by Spot could be followed up by maintenance technicians.

By mid-April, 23 intervention requests had already been registered: 18 for air leaks and 5 for acoustic anomalies on motors.

This illustrates the direct link between robotic inspection and actual maintenance action on the shop floor.

The project was not limited to maintenance. Spot was also configured for safety and security missions.

The robot can, among other things, check the presence and condition of safety equipment such as fire extinguishers, first-aid kits and hose reels. Spot was also used for night patrols. During these rounds, the robot can detect the presence of someone on site and send an alarm with an image.

Spot therefore takes on a broader role, becoming a multifunctional tool for maintenance, safety and site security.

The results are significant. At the start of Léo's project, Spot had a limited number of missions and analyses. Following the development phase, the robot evolved from approximately 8 missions and 300 analyses to around 18 active missions and more than 700 regularly performed analyses.

The use of Spot was therefore significantly expanded and diversified. The robot can now perform thermal and acoustic analyses and check meters on critical equipment across the site.

The project also demonstrated that robotics is not simply a matter of installing technology.

Spot does not automatically “understand” a factory. The robot needs to be carefully programmed, routes need to be adapted, analysis zones need to be correctly defined and thresholds need to be validated.

Léo therefore also worked on reducing false alarms, improving the readability of pressure gauges and adapting analyses to the actual conditions on site. This required taking into account factors such as insufficient lighting, obstacles, motorised shutters, pedestrian crossings, stairs and slippery areas.

The robot itself also requires maintenance. Léo therefore developed a maintenance sheet for Spot, covering aspects such as cleaning, fans, cameras, equipment mounted on the robot and the contact points of the docking station.

When a robot becomes part of the inspection process, its own availability also becomes a maintenance issue.

During the evaluation by the BEMAS jury, Léo's internship supervisor highlighted his independence, commitment and maturity. He also emphasised the time Léo spent on the shop floor, the importance of understanding machines and potential failure modes, and his ability to work with maintenance and IT teams, automation specialists, suppliers and Boston Dynamics.

For BEMAS, this project is representative of the evolution of the maintenance profession.

Robotics, artificial intelligence, data collection and CMMS are not objectives in themselves. Their value lies in their ability to help teams better observe equipment, understand anomalies and prioritise maintenance interventions more effectively.

Léo Chastain's work demonstrates how a young technician can contribute concretely to this transformation by combining electromechanical knowledge, a practical mindset, technological curiosity and field expertise.

“This project perfectly illustrates the new face of industrial maintenance.”
“Technologies such as Spot do not replace maintenance technicians. They give them new means to detect anomalies earlier, structure data and intervene more effectively. These are precisely the kinds of skills our industry will need in the years to come.”

Wim Van Cauwenberghe – Director of BEMAS

The project opens up several interesting perspectives. Full automation of the transfer of anomalies to the CMMS, additional AI capabilities, an expansion of the missions and the use of historical data could contribute to a more advanced condition-based maintenance approach.

The use of Spot for remote acoustic or ultrasonic analyses of certain motors is also a promising avenue, provided that the quality of the data and its relevance for the early detection of faults are carefully validated.

By highlighting this final-year project as part of the Professor Leenaerts Award, BEMAS aims to emphasise the importance of practice-oriented education, applied innovation and young talent in maintenance, reliability and asset management.

Léo Chastain's work demonstrates that Industry 4.0 technologies only truly prove their value when they serve a clear purpose: improving control of equipment, supporting maintenance technicians and strengthening industrial performance.