XTRACTIS FOR predictive maintenance

Prediction of the Degradation of a Naval Propulsion Unit

Benchmark vs. Random Forests, Boosted Trees & Neural Networks

Design an AI-based decision system that accurately predicts the functional degradation of a naval propulsion unit compressor, given the hyper-complexity of the phenomenon (strongly nonlinear behavior) in order to rationally plan explainable maintenance operations.
Goals & benefits

Allow business experts and maintenance managers to understand the causal relationships between some turbine parameters and its future state of degradation.

Find the truly influential parameters for assessing the state of degradation and thus reduce measurement and maintenance costs.

Carry out maintenance actions specific to each turbine upstream in order to avoid critical damage, thanks to rapid and systematic diagnostics, while justifying each intervention.

  •  The top-model is a decision system composed of 428 gradual rules without chaining.
  • Each rule uses from 1 to 10 predictors among the 12 variables that XTRACTIS identified as significant (out of the 14 turbine parameters).
  • The model is relatively intelligible despite the large number of rules, given the high complexity of the studied phenomenon.
  • Only a few rules are triggered at a time to compute the decision

It has an excellent Real Performance (on unknown data).

It computes real-time predictions up to 70,000 decisions/second, offline or online (API).

UC02 scores graph

RFo=Random Forests
BT=Boosted Trees
NN=Neural Networks

Detailed results and explanations in full document

Use Case 2024/03 (v6.1)

Powered by XTRACTIS® REVEAL 12.2.43064 (2022/08)


  1. Problem Definition
  2. XTRACTIS-induced Decision System
  3. XTRACTIS Process
  4. Top-Model Induction
  5. Explained Predictions for 2 unkown cases
  6. Top-Models Benchmark
  7. Appendices