Introduction

The invention is an anomaly detection and prescriptive maintenance system which, based on the measurements of various sensors and various artificial intelligence logics, is able to identify critical events in infrastructures (buildings, telecommunications pylons, bridges, etc..) enabling the maintainers to check and restore them.

Technical features

The invention consists of a system able to identify critical events that could have impact on infrastructures (buildings, telecommunications pylons, bridges, etc.) and provide alerts to domain experts, allowing the system to be adapted to different types of infrastructures, with different ages and with different uses. The system is composed of three main logics: (1) a predictive mechanism to detect local anomalies, (2) a process to identify if the set of sensors is describing a global anomaly of the structure, and finally (3) a learning system that determine the optimal decision to verify the infrastructure by considering the prediction about the structure and the feedback of the maintainers.

Possible Applications

• Civil constructions
• Telecommunications pylons
• Railways and roads
• Industrial installations
• School buildings

Advantages

• Optimization of maintenance operations
• Versatility with respect to different contexts
• Adaptability of decision-making strategies considering the feedback given by experts
• Independence from the sensors used
• Continuous accessibility to the state of well-being of the structure