CASE Study 1

Prognosis of remaining useful life of aircraft turbofan engine

The training data contains simulated run-to failure data for different airline turbofan engines. Engine degradation simulation was carried out using C-MAPSS by the Prognostics CoE at NASA Ames. Four different data sets simulated under different combinations of operational conditions and fault modes. The simulation records several sensor channels to characterize fault evolution. All datasets had 26 inputs comprising of sensor data and operational settings.

Prognostics in this case study is defined as the estimation of remaining useful component life in units of time (e.g. hours or cycles). The dataset had common challenges in the prognostic field such as noise from sensor data, and different simulated failure modes. The inclusion of noise, and machine initial wear ensured that the data simulated real-life scenarios.

Our AI solution was able to bring down the error value to just 12.07.


The NASA CMAPSS data includes four different datasets: FD001-FD004. They are different based on the operational conditions and the fault modes.

 There are 3 operational settings in the simulations

  • Altitude
  • Mach number
  • Sea level temperature

 Each operational mode is a combination of these operational settings. The maximum number of operational modes, as can be seen from the table, is six.


The bar chart shows the comparison of RMSE obtained using Fuzzy Bolt against Neural Nets.