Start-Ups Bring AI Tools To Predictive Maintenance

Start-up companies make their presence known in big data analytics designed for predictive maintenance functions.

The new era of Big Data and sophisticated analytics for predictive maintenance has drawn the interest of some famous and very big companies. It has also attracted the attention of small, start-up firms that, in so many other areas of the digital revolution, have done things faster or better than the giants could. Innovative Binaries is one such software start-up, recently selected as one of seven start-ups among 137 applicants by Airbus for its BizLabs program.

Innovative Fonder Soumitra Miraj says his analytic platform is still under development, but some small parts of it are already deployed. “The algorithms are being applied to a wide variety of aircraft, turboprops, helicopters, business jets and narrowbody jets,” Miraj says. “We are addressing several main systems and subsystems within the aircraft, nose to tail.” As with other analytic tools, the coverage of Innovative’s solution depends on the sophistication of the aircraft and the type and quality of data generated by sensors. 

Distinctively, Innovative has already partnered with at least one long-time MRO management software provider, CALM Systems. Miraj says he partnered with four companies in all and “a couple more are likely to sign up quite soon.”

How good will Innovative Binaries' predictive maintenance tools be? The Innovative team has expertise in a number of artificial intelligence domains: neural networks; genetic algorithms; and both supervised and unsupervised machine learning. “Because aircraft are so reliable, it is almost impossible to obtain data samples for every type of fault,” Miraj notes. “The most useful approach is to learn from the mass of healthy data in order to detect abnormalities.” He believes in the value of supervised learning, in which huge amounts of data from systems known to be healthy are analyzed so anomalies or departures from healthy patterns can be detected. “The key is to offer early warning of any impending problems. The best way to do that is to quantify aircraft health, so small changes can immediately be flagged.”

The Innovative exec thinks his independent experts have some advantages over alternative analytics providers. “We do not have any OEM agenda. Our platform is built for aircraft operators to help them effectively use the data they generate. The system significantly increases their decision-making capabilities and negotiating muscle.”

In addition, Miraj says many OEM analytic tools are decades old, not designed for massive data analysis and often not flexible enough to incorporate machine learning. He acknowledges these drawbacks to OEM systems are ending “slowly.” But Innovative now has several years of experience in applying modern machine learning to actual aircraft data.

In addition to data analytics for predictive maintenance, Innovative also offers inventory optimization tools. These are just in the test stage right now, but Miraj believes inventory optimization will be highly useful. “The solution brings a new perspective to planning, stocking and optimizing inventory. It provides a new intelligent layer, wrapped around existing rules and policies, that continuously evaluates hundreds of parameters before making recommendations.” And Innovative’s inventory tools are tightly integrated with its predictive maintenance platform.

Given the massive amount of data generated by modern aircraft and useful for predictive maintenance, Miraj says data will generally be stored in the cloud. But he has “different engagement models for very large operators.”

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.