Dynamic Weighted PSVR-Based Ensembles for Prognostics of Nuclear Components

Jie Liu (1) , Valeria Vitelli (2) , Redouane Seraoui (3) , Enrico Zio (1)

1 SSEC - Chaire Sciences des Systèmes et Défis Energétiques EDF/ECP/Supélec
2 Department of Biostatistics
3 EDF R&D - EDF R&D

Combining different physical and / or statistical predictive algorithms for Nuclear Power Plant (NPP) components into an ensemble can improve the robustness and accuracy of the prediction. In this paper, an ensemble approach is proposed for prediction of time series data based on a modified Probabilistic Support Vector Regression (PSVR) algorithm. We propose a modified Radial Basis Function (RBF) as kernel function to tackle time series data and two strategies to build diverse sub-models of the ensemble. A simple but effective strategy is used to combine the results from sub-models built with PSVR, giving the ensemble prediction results. A real case study on a power production component is presented.

Domaines: Energie électrique; Sciences de l'ingénieur [physics]; Machine Learning [stat.ML]

Jie Liu, Valeria Vitelli, Redouane Seraoui, Enrico Zio. Dynamic Weighted PSVR-Based Ensembles for Prognostics of Nuclear Components. FLINS 2014, Aug 2014, João Pessoa, Brazil. ⟨hal-01108176⟩