State Estimation in Solid Oxide Fuel Cell(SOFC)System

Polisetty, V G and Jampana, Phanindra Varma (2013) State Estimation in Solid Oxide Fuel Cell(SOFC)System. Masters thesis, Indian Institute of Technology, Hyderabad.

Preview

Text
CH11M1015.pdf - Submitted Version
Download (2MB)

Abstract

Solid oxide fuel cells (SOFCs) oer a clean, low pollution technology to electrochemically generate electricity at high eciency. An SOFC consists of a dense solid electrolyte and two porous electrodes in contact with an interconnect on either side. The control of an SOFC stack becomes important in order to ensure adequate and disturbance free electric power. As several controlled/constrained variables are not directly measured in a stack, state estimators can be used in order to study the dynamic behaviour of SOFC stacks as well as to design eective SOFC controllers. In this thesis, A zero dimensional model represented by a set of ordinary dierential equations is derived for dynamic modeling. The model consists of molar balances and an energy balance coupled with a simplied description of the fuel cell electrochemistry. The chemical species considered are H2 and H2O for fuel side (anode side) and O2 and N2 for air side (cathode side) and the electrochemical model accounts for ohmic, concentration and activation losses. Considering the estimation part, the state vector which is to be estimated consists of partial pressure of chemical species and temperature, with voltage as the measurement. Estimation of states for linear systems can be done by Kalman Filter. States of nonlinear systems can be estimated using Extended Kalman Filter(EKF), Unscented Kalman Filter (UKF). We choose UKF for non linear state estimation. UKF is a derivative free state estimator for non linear systems. This work investigates the use of non linear state estimator UKF to estimate the states of SOFC system. This method can be applied to estimate states in any type of fuel cells (PEMFC, AFC etc.) by very slight modications.

[error in script]

IITH Creators: