Is it possible to early detect faults in a manufacturing production line?
In our latest publication ‘Divide, Propagate and Conquer: Splitting a Complex Diagnosis Problem for Early Detection of Faults in a Manufacturing Production Line’, Kerman Lopez de Calle – Etxabe, Meritxell Gomez – Omella and Eider Garate – Perez elaborate on their solution of the data challenge proposed by the 6th European Conference of Prognostics and Health Management Society 2021 (PHM). This challenge deals with a manufacturing line that continuously tests fuses and suffers from several malfunctions.
The proposed solution addresses:
- the diagnosis of the faults;
- the efficiency of the diagnosis;
- the identification of the signals related to each fault type;
- the identification of different operation settings that occur during the non-faulty conditions.
As the authors explain, this problem presents some difficulties that are common to machine fault diagnosis or manufacturing line monitoring, such as the class imbalance, the high amount of missing values, multicollinearity and high dimensionality and experimental noise. Additionally, they point out that the evaluation criteria present further challenges, such as the consideration of chronology and the detection of operation states.
Furthermore, the consideration of all these factors turned this exercise in a very representative and challenging problem. In specific, this proposed solution that has actually obtained the highest score in the contest, relies on the combination of decision tree algorithms and a propagation system. As it is highlighted, the decision tree algorithms provide observation-wise diagnoses, while the propagation system deals with chronology by adding a Kalman style filter that updates the probabilities, resulting in a more reliable result.
You can read the publication online here.