/pred_maintenance

Predictive maintenance of an aircraft engine

Primary LanguageJupyter NotebookMIT LicenseMIT

Predicting the remaining uselful life of a turbofan engine

September 2019

Introduction

Project Description:

We will use this dataset to analyse and predict a the remaining uselful lifespan (RUL) of a turbofan engine.

Data description:

Data stems from the data challenge competition held at the 1st international conference on Prognostics and Health Management (PHM08) is being made publicly available. It contains recorded sensor data of 21 sensors in 218 engines. Turbofans were operated for multiple cycles and with different settings. The data set was provided by the Prognostics CoE at NASA Ames.

Source: A. Saxena and K. Goebel (2008). "PHM08 Challenge Data Set", NASA Ames Prognostics Data Repository (http://ti.arc.nasa.gov/project/prognostic-data-repository), NASA Ames Research Center, Moffett Field, CA

Background

The remaining uselful life (RUL) denotes the lifespan of an operating system from its current condition until failure. This can be expressed in time (e.g. hours), miles, cycles etc. There are generally three approaches to predict the RUL, depending of which data is existent from previous failures of the same system type (other similar turbofan engines):

  1. Survival model, if only data from the failure of previous systems exists,
  2. Degradation model, if thresholds define the safe operation of the system,
  3. Similarity model, if continous data about of the operation of the system (including its degradation) exsits (run-to-failure data).

The Survival model uses the probability of the occurance of failure in previous systems to estimate the RUL. The Degradation model uses the past information to predict when the system will pass the safety threshold. Finally, the similarity model uses the histories of previous run-to-failure data for predicting RUL.

Approach

Results and Conclusions