Computational Intelligence Driven Prognostics for Remaining Service Life of Power Equipment

Abstract

Maintenance of power equipment is generally based on fault occurrence since most of the diagnostic tests and assessment procedures come into play once the asset or component failure is reported. Estimation of remaining useful (RUL) life can help utilizing the devices to its maximum and can increase life because of in-time maintenance. A brief comparison of shallow and deep learning techniques leads to the choice of deep architectures for prognostic of RUL because of the nature of multimodal data.