The Lisa Group is a year-and-a-half old startup with a staff who have plenty of experience in using artificial intelligence to do predictive maintenance, according to managing director Ali Baghchehsara. The company has some new ideas on how predictions should be done.
“We take a different approach to predictive maintenance,” says Baghchehsara, who has been working on predictive maintenance for a variety of industries for about eight years.
OEM platforms like Airbus’s Skywise analyze massive volumes of data and then engineers manually write predictive algorithms, the Lisa exec explains. In contrast, Lisa analysts use robots to write or select the algorithms that work best with the data.
Lisa tested this approach with a demonstration project at a major airline. At first, the robots were not finding very good predictive algorithms. So analysts tweaked a few things and the predictions got a lot better.
The 16 Lisa staffers were thus able to produce predictions that took better account of special events like hard landings than the conventional predictive techniques, which Baghchehsara calls “essentially, regression.”
For the airline demonstration Lisa applied its robotic approach to heating, ventilation and air-conditioning, landing gear and water waste—or toilet—systems. In June, Lisa and the client airline will jointly present results of the project at a London conference on predictive maintenance.
Much more may be ahead. Next will come application to cabin systems. By 2025, Baghchehsara expects to have a general predictive system, complete with visual dashboards, that is applicable to any aircraft system.