United Technologies Corp.’s [UTX] recent acquisition of Predikto will provide military maintenance technicians with improvements on their ability to predict when unscheduled events will occur within F-35, F-22, F-16 and eventually KC-46A engines.
The acquisition will integrate Predikto’s predictive analytics software across UTC’s entire portfolio, including aftermarket support for in-service military aircraft engines manufactured by subsidiary Pratt & Whitney.
Predikto, an Atlanta-based provider of predictive data analytics, was acquired by UTC in early August under undisclosed terms. The acquisition is finalized, and will expand Predikto’s presence from its Atlanta headquarters to UTC’s digital accelerator facility in New York City, a complex dedicated to researching and developing new digital approaches to providing aftermarket data analytics support to in-service aircraft engines.
Under the acquisition, Mario Montag, former CEO of Predikto, has become the new chief data and analytics officer for UTC. Ahead of the acquisition becoming final, Predikto completed pilot programs involving Pratt & Whitney commercial engines and used a combination of pre-existing operational history and existing physics-based models to identify factors related to unplanned engine events. UTC reports that the software provided by Predikto was able to aggregate the data and create algorithms that “only took days” for a process that usually takes engineers several months to complete.
“Predikto operates on the data captured – we bring no sensors, no hardware requirements – we are laser focused on delivering actionable results quickly and at scale. This is in addition to the services already offered such as EngineWise, the suite of products offered by Pratt & Whitney for large commercial engines, and the ongoing sustainment offering to military customers,” Montag said.
Montag describes the technology as automating the historically manual and time intensive modeling processes of feature engineering, feature selection, and the training and evaluation of machine learning models. Predikto’s platform will also be capable of ingesting geospatial and atmospheric data about the operational environments of military aircraft engines supplied by Pratt & Whitney.
“As the Predikto platform learns those pieces of data most informative for predicting the events of interest, this information can be used by to evaluate the usefulness of different sensors, sometimes providing indication of where to invest more in digitization,” Montag said.