Title: Proton exchange membrane fuel cell dynamic model based on time series analysis for fault diagnosis
Authors: Sujit Sopan Barhate; Rohini Mudhalwadkar
Addresses: Department of Technology, Savitribai Phule Pune University, Pune – 411007, India ' Instrumentation and Control Department, College of Engineering, Pune – 411005, India
Abstract: Proton exchange membrane fuel cell is being used in automobiles as it is a clean energy generator. However, there is a need to improve its useful life in the automotive application. Fuel cell life can be improved by detecting fault conditions and correcting them in run time. Fault detection needs a robust model to predict fuel cell output. In this paper, we propose an autoregressive moving average (2, 1) model for the fuel cell. The model is experimentally validated in static as well as dynamic conditions. The mean square error of the model predicted output is less than 2 × 10−5. A novel approach for fault detection and isolation is proposed and experimentally validated. Fault in the fuel cell is detected by comparing the change in cell power per change in cell voltage and change in cell voltage per change in cell current of the cell under operation with the healthy cell.
Keywords: proton exchange membrane; PEM fuel cell model; fuel cell faults; fuel cell faults isolation; ARMA model.
International Journal of Renewable Energy Technology, 2021 Vol.12 No.4, pp.351 - 379
Received: 10 Dec 2020
Accepted: 11 Mar 2021
Published online: 02 Sep 2021 *