The project

Production systems in manufacturing and mining are constantly under stress, due to the need to meet production goals and to reduce the costs. Through the current context of increasing usage of sensors in monitoring of production systems, alongside the fact that prognostics is one of the main development areas worldwide, the development of “Prognostics and Health Management” (PHM) techniques is becoming more and more attractive for companies to incorporate them into their maintenance systems.

Developing an appropriate PHM system allows the failures to be anticipated in critical systems or components. Such a system prevents risks by increasing the safety of people and goods, the reliability and the interval between maintenance operations. Besides, it reduces emergencies associated with unforeseen failures; decreases maintenance costs; and improves the quality of production. Moreover, the damage prognostics is aligned with the principle of sustainability, which means an increase in the availability and useful life of the systems.

Currently, most of the equipment and systems are being monitored continuously, generating a huge amount of data (Big Data). This implies a great opportunity to improve the reliability and maintenance techniques of a structure or an equipment. In the Department of Mechanical Engineering at University of Chile, PHM techniques are being developed based on the combination of deep learning tools and the fault physics, to be used in monitoring and health prognostics of machinery or structures under uncertainty. Particularly, Different algorithms have been developed to detect structural damage from images, process sensors (such as pressure, temperature, flow), vibration monitoring, as well as the detection and prognostics of failures in rotating machines from monitoring and fusion of various types of sensors.

Our technological development will allow the industry to:
  • Evaluate the health of a set of machines using a single method, instead of developing a customized approach for each equipment variant.

  • Develop a monitoring system that is robust against changes in the operating environment, since today’s modern structures operate under different conditions of speed, load and temperature.

  • Minimize the dependence on the availability of labeled training data.

In this regard, we have the appropriate infrastructure through the Smart Reliability and Maintenance Integration Laboratory (SRMI Lab), which is focused mainly on R & D and consultancy in the area of deep learning, Big Data and IoT; applied to the diagnostics and prognostics of machinery failures.

SRMI Lab (


An example of our work in the laboratory is the development of a web application called Deep Learning Hub ( which aims to help the graphic design of artificial intelligence algorithms, without having previous knowledge of programming.



Dr. Enrique López
Project Director 

Dra. Viviana Meruane
Alternate Director

Back to Projects