INESC TEC Science Bits – Episode 32
Rita Ribeiro, INESC TEC researcher and professor at the Faculty of Sciences of the University of Porto
Keywords: fault detection and diagnosis, predictive maintenance, explainable AI
Looking for rare events, faults and malfunctions
If we travel back in the history of maintenance, it’s easy to understand how centuries of evolution separate reaction from proactivity. Over the years, the industrial progress led to the development of solutions to address – or to anticipate – equipment malfunctions, in order to avoid losses. In 1919, the manual of Ford’s T model asked clients to take care of their vehicles and pay special attention – brief actions that could prevent delays or accidents.
The maintenance operations evolved, and the emergence of Industry 4.0 led to the development of technologies to detect issues even before they happen. How? Through data analysis, pattern identification and detection of malfunction. All these thanks to large volumes of data and AI solutions.
Basically, we’re talking about predictive maintenance. In this sense, we invited Rita Ribeiro, INESC TEC researcher, to tell us how the Institute (in partnership with Metro do Porto) developed a solution focused on the detection of faulty vehicles, thus reducing maintenance costs. We also talked about explainable AI in predictive maintenance. Join us in another episode of Science Bits.