Aircraft Predictive Maintenance Software
A software package based on a predictive algorithm makes it possible to estimate the residual life of aircraft components and assemblies with a relatively high accuracy and plan their maintenance and repair.

Currently, aircraft technicians are planning to repair aircraft and helicopters either based on the detected problems or on the basis of recommendations from aircraft manufacturers.
It is assumed that thanks to the predictive algorithm, the repair and maintenance of machines can be carried out as needed. As a result, technical services will be able to reduce the number of operations carried out with aircraft, which will ultimately lead to a reduction in the downtime of aviation equipment.
The new system involves the installation of a variety of sensors on aircraft that monitor the operation of onboard systems, components and assemblies of aircraft. The recorded data from the sensors is then used for analysis for possible hidden problems and for predicting the time of occurrence of failures. Based on the analysis, schedules for the repair and maintenance of aircraft are compiled. Also, the new aircraft maintenance system will allow saving on the order of new and repair of old units. In particular, thanks to it, it will be possible to replace parts according to their actual condition, and not according to manufacturers' recommendations.
Product for preventive replacement/repair of aircraft components
based on predictive analytics allows
Reduce financial and reputational losses by reducing unscheduled aircraft downtime for technical reasons
    Reduce the cost of replacing or repairing expensive components through early detection of pre-failure conditions
    Increase revenue by reducing aircraft downtime during scheduled maintenance
    The effect of the implementation
    predictive maintenance

    The implementation of the solution would reduce downtime due to technical reasons in 2017 by 460 hours per year, or by 52% for the Airbus fleet alone.

    Approach to forecasting

    The developed solution predicts the occurrence of an event signaling the failure of an aircraft component for a certain number of flight cycles ahead.

    Model Quality Assessment

    The target variable is a binary indicator of the occurrence of at least one event in the forecast window.
    Calculation of the quality of work of models is carried out on the target vector.

    Target business process

    The developed solution makes it possible to predict the onset of 5 types of defects in engines and a defect in the hydraulic system.
    In the future, it is planned to expand the amount of data received from the on-board sensors of the aircraft and build predictive models for other types of defects.

    Architecture of the Predictive maintenance software package

    User interface
    Monitoring of aircraft parameters
    The solution uses data from the following sources
    • 1
      Indications of onboard sensors and messages from the aircraft information system
    • 2
      Data on maintenance and flight of parts from the accounting system of the airline
    • 3
      Flight schedule data from the airline's accounting system
    • 4
      Data on weather conditions at the airports of the route network
    We have been developing advanced solutions based on neural networks since 2016
    The economic effect of cooperation with us -
    tens of thousands of dollars a day
    Our contacts
    +7 (967) 215-75-05