Predicting how long a machine will keep running before it fails - its remaining useful life - is the core problem of predictive maintenance, but real run-to- failure data is scarce and expensive. To give researchers a common testbed, NASA released the C-MAPSS turbofan engine degradation dataset, introduced in the paper “Damage Propagation Modeling for Aircraft Engine Run-to-Failure Simulation” by Saxena, Goebel, Simon, and Eklund at the 2008 International Conference on Prognostics and Health Management.
The data was generated with NASA’s Commercial Modular Aero-Propulsion System Simulation (C-MAPSS), which models a turbofan engine and lets researchers inject gradual damage and record many sensor channels as the fault evolves. The release contains four sub-datasets (FD001 through FD004) covering different operating conditions and fault modes, each split into training engines run to failure and test engines stopped partway, where the task is to predict how many operating cycles each test engine has left.
C-MAPSS became one of the most widely used benchmarks in the prognostics and health management community, and matters to a general reader as the shared yardstick that let predictive-maintenance methods - now used on jet engines, turbines, and factory machinery - be compared on equal footing.