Constrained model predictive control for proton exchange membrane fuel cell

A constrained model predictive control (MPC) is designed to regulate the air flow rate of proton exchange membrane fuel cell (PEMFC). Oxygen excess ratio, compressor flow rate and supply manifold pressure are constrained to avoid oxygen starvation, surge and choke phenomena. This is achieved by ma...

Full description

Saved in:
Bibliographic Details
Main Authors: Abdullah, Muhammad Faris, Idres, Moumen
Format: Article
Language:English
English
Published: Korean Society of Mechanical Engineers 2014
Subjects:
Online Access:http://irep.iium.edu.my/43832/1/2014-Constrained_model_predictive_control_of_proton_exchange_membrane_fuel_cell_2014.pdf
http://irep.iium.edu.my/43832/4/43832_Constrained%20model%20predictive_SCOPUS.pdf
http://irep.iium.edu.my/43832/
https://link.springer.com/content/pdf/10.1007%2Fs12206-014-0849-0.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A constrained model predictive control (MPC) is designed to regulate the air flow rate of proton exchange membrane fuel cell (PEMFC). Oxygen excess ratio, compressor flow rate and supply manifold pressure are constrained to avoid oxygen starvation, surge and choke phenomena. This is achieved by manipulating compressor voltage and stack current. The choice of the manipulated input to satisfy a constraint is investigated. Surge and choke avoidance is successful, when compressor voltage is manipulated. When stack current is utilized to satisfy surge and choke constraints, a large unrealistic current is needed. Oxygen starvation is successfully avoided utilizing stack current, while compressor voltage manipulation fails to prevent oxygen starvation. Thus, a current governor is implemented to handle oxygen starvation, while the compressor voltage is constrained to avoid surge and choke. Quadratic programming optimization, Laguerre and exponential weight function are employed to reduce the computational burden of the controller. The simulation results prove that the proposed controller managed to satisfy all constraints without any conflict.