Project description

Optimize maintenance policies for production systems based on forecast maintenance

With the massive introduction of control mechanisms and components integrating software parts, industrial production systems are becoming more complex. One of the major challenges today is to improve the methods and tools used to help keep these systems operational. Predictive maintenance, coupled with the optimization of maintenance policies, represents a strong industrial interest because it allows for the better planning of maintenance and repair operations, the reduction of the risks of breakdowns and unplanned shutdowns, and the increase of the equipment’s life span.

Launched in November 2018 for a period of 4 years, DFO project aims to overcome the technological and methodological barriers of predictive maintenance and the combination of maintenance policies in production systems, made possible by new technologies and artificial intelligence, and the computing power of the machines, in order to optimize their maintenance in operational condition.

Implemented skills

	Data science and AI
	Optimization
	Dependability of critical systems

Targeted targets

• Industrial processors and manufacturers with a large fleet of systems

Supervised thesis in the framework of the project

Thesis #1: Deep learning and business knowledge for the monitoring and diagnosis of failures from massive complex data (CentraleSupélec / IRT SystemX)
Thesis #2: Dynamic grouping of maintenance tasks for multi-component systems with multiple degradation modes (Ifsttar / IRT SystemX)