Modeling of a Hot Strip Mill Temperature using Hybrid Learning Interval Type-1 and Type-2 Non-singleton Type-2 FLSs

M. Méndez, A. Cavazos, L. Leduc, and R. Soto (Mexico)


Interval type-2, stationary noise, non-stationary noise, temperature prediction, hot strip mill, least-squared, back propagation, square-root filter, Mamdani ANFIS type-2.


This paper presents results of a research on two proposed hybrid learning algorithms: back-propagation (BP) recursive least-squared (RLS) and back-propagation (BP) square-root filter (REFIL) - for interval singleton type-2 Fuzzy Logic System (FLS) (type-2 SFLS) parameters estimation [1]. The algorithms are extended to interval type-1 non-singleton type-2 FLS (type-2 NSFLS-1) and interval type-2 non-singleton type-2 FLS (type-2 NSFLS 2). The proposed algorithms are evaluated in a Hot Strip Mill (HSM) Scale Breaker (SB) entry surface temperature prediction application. A comparative result shows the advantage of hybrid learning methods over that with only BP [2].

Important Links:

Go Back