TY - GEN

T1 - Case Study

T2 - 3rd European Conference on Electrical Engineering and Computer Science - EECS 2019

AU - Santos Neto, Accacio Ferreira dos

AU - Santos, Murillo Ferreira dos

AU - Santiago, Augusto Cerqueira

AU - Vidal, Vinícius Ferreira

AU - Mercorelli, Paolo

N1 - The conference was held in December 2019 and the paper appeared in IEEEXplore and in Scopus in December 2020.

PY - 2019

Y1 - 2019

N2 - This work proposes the identification study of a small scale nonlinear system through Neural Network AutoRegressive with eXternal input (NNARX). The use of the NNARX network requires a proper methodology for its configuration and a good training set. This work proposes the main parameter definition of the network through the analysis of non-intrusive performance indexes with the support of a database based on the system response, considering a Pseudo-Random Binary Sequence (PRBS) excitation. The methodology will be applied in two specific open-loop identification situations: Numerical simulation of a fourth order polynomial system (Case 01), and an experimental system that controls the water tank level with nonlinear characteristics (Case 02). The results of the identified models by the developed NNARX network were able to represent the system dynamics with high fidelity, presenting an average identification error of less than 0.14% (Case 01) and 0.34% (Case 02). It was also observed the learning and generalization evidence that could represent the process intrinsic nonlinearities satisfactorily. Besides, it will be possible to find the potentiality and usefulness of the NNARX network in non-linear system identification. Besides, it will be possible to find the potentiality and usefulness of the NNARX network in nonlinear system identification.

AB - This work proposes the identification study of a small scale nonlinear system through Neural Network AutoRegressive with eXternal input (NNARX). The use of the NNARX network requires a proper methodology for its configuration and a good training set. This work proposes the main parameter definition of the network through the analysis of non-intrusive performance indexes with the support of a database based on the system response, considering a Pseudo-Random Binary Sequence (PRBS) excitation. The methodology will be applied in two specific open-loop identification situations: Numerical simulation of a fourth order polynomial system (Case 01), and an experimental system that controls the water tank level with nonlinear characteristics (Case 02). The results of the identified models by the developed NNARX network were able to represent the system dynamics with high fidelity, presenting an average identification error of less than 0.14% (Case 01) and 0.34% (Case 02). It was also observed the learning and generalization evidence that could represent the process intrinsic nonlinearities satisfactorily. Besides, it will be possible to find the potentiality and usefulness of the NNARX network in non-linear system identification. Besides, it will be possible to find the potentiality and usefulness of the NNARX network in nonlinear system identification.

KW - Level System

KW - NNARX networks

KW - System Identification

KW - Engineering

UR - http://www.scopus.com/inward/record.url?scp=85097936758&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/d5ca1f0e-1c31-387b-9b55-dfa7c2537b6d/

U2 - 10.1109/EECS49779.2019.00016

DO - 10.1109/EECS49779.2019.00016

M3 - Conference contribution

AN - SCOPUS:85097936758

SN - 978-1-7281-6110-5

SP - 15

EP - 18

BT - 2019 3rd European Conference on Electrical Engineering and Computer Science, EECS 2019

PB - IEEE - Institute of Electrical and Electronics Engineers Inc.

CY - Piscataway

Y2 - 28 December 2019 through 30 December 2019

ER -