PROJECT TITLE :

Comprehensive Modeling of U-Tube Steam Generators Using Extreme Learning Machines

ABSTRACT:

This paper proposes artificial neural network and fuzzy system-based mostly extreme learning machines (ELM) for offline and online modeling of U-tube steam generators (UTSG). Water level of UTSG systems is predicted in a very one-step-ahead fashion using nonlinear autoregressive with exogenous input (NARX) topology. Modeling information are generated employing a well-known and widely accepted dynamic model reported within the literature. Model performances are analyzed with totally different range of neurons for the neural network and with completely different number of rules for the fuzzy system. UTSG models are designed at completely different reactor power levels in addition to full range that corresponds to all or any reactor operating powers. A quantitative comparison of the models are created using the basis-mean-squared error (RMSE) and also the minimum-descriptive-length (MDL) criteria. Furthermore, standard back propagation learning-primarily based neural and fuzzy models are also designed for comparing ELMs to classical artificial models. The benefits and drawbacks of the designed models are discussed.