Skip to Main Content
Condition Based Maintenance (CBM) software, called cbmLAD, under development at École Polytechnique de Montréal is presented in this paper. The backbone of the software is a supervised learning data mining approach called Logical Analysis of Data (LAD). LAD possesses distinctive advantages that are useful in Condition Based Maintenance (CBM), namely its independence from statistical processes and its ability to generate interpretable patterns. The latter property serves to reinforce the theoretical knowledge and uncover new knowledge about a certain diagnostic problem in CBM. cbmLAD has been tested in two maintenance scenarios. Expert knowledge was elicited in each scenario to train the diagnostic decision models obtained through cbmLAD. This paper describes the methodology applied in each scenario and highlights the advantages of using LAD for fault diagnosis.