Premium
Subspace‐based Wavelet preprocessed Data‐driven Predictive Control
Author(s) -
Vajpayee Vineet,
Mukhopadhyay Siddhartha,
Tiwari Akhilanand Pati
Publication year - 2016
Publication title -
incose international symposium
Language(s) - English
Resource type - Journals
ISSN - 2334-5837
DOI - 10.1002/j.2334-5837.2016.00337.x
Subject(s) - subspace topology , wavelet , preprocessor , noise (video) , computer science , controller (irrigation) , model predictive control , process (computing) , data pre processing , control theory (sociology) , artificial intelligence , data mining , pattern recognition (psychology) , control (management) , agronomy , image (mathematics) , biology , operating system
This paper introduces a methodology for designing a subspace‐based data‐driven predictive control with wavelet preprocessing. In a data‐driven control, especially when SNR is low, it becomes difficult to obtain reliable predictor coefficients. Therefore, it is imperative to have a processed and informative dataset for stable controller operation. Wavelet being capable of better noise rejection from process dynamics motivates to perform wavelet filtering before designing the control law. Methodology for deriving the predictor from subspace matrices of processed data is presented. A predictive controller, estimated from the dataset, is designed for power control of a nuclear reactor core for a load‐following operation. The efficacy of the proposed technique is demonstrated by Monte Carlo simulations in stationary as well as non‐stationary noise cases.