Research on Prognostic Method Based on SVR Optimized by Improved QPSO
Year: 2014
Type: article
Abstract: Prognostic capability is a very important character for PHM distinguishing from other diagnosis systems, fault prognostic method based on SVR was researched on in this paper. In order to solve kernel function parameters, penalty coefficient and insensitivity loss coefficient of SVR, improved QPSO was put forward to train SVR in this paper. And then a certain power supply combination in beam control system of a certain control and guide radar was taken as an example to collect voltage signal and forecasted, and its precision was very perfect and could satisfy prognostic demand.
Source: Applied Mechanics and Materials
Authors: Wan Ling Li, Xiang Jun Song, Yan Chen
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W2115676376