Modeling and predicting the membrane water content of proton exchange membrane fuel cell by using support vector regression

摘要

This study examines the use of the support vector regression (SVR) approach in modeling and predicting the membrane water content of Proton Exchange Membrane Fuel Cell (PEMFC) under two influence factors, including the impedance of single-PEM-chip and operating temperature. The leave-one-out cross validation (LOOCV) test results by SVR strongly support that the generalization ability of SVR model is high enough: mean absolute error (MAE) is 0.01, mean absolute percentage error (MAPE) is 0.15% and multiple correlation coefficients (R~2) is 1.00. This investigation suggests that the SVR approach is a promising and practical methodology to simulate the properties of fuel cell system.