Functional characterization of current characteristic of direct methanol fuel cell

In the last few years, direct methanol fuel cells (DMFCs) are extensively considered as a viable alternative power source to batteries used in transportation and portable devices due to their high energy density properties. Despite these significant advantages, operational and technological improvements are still required to make them a cost effective process. Past studies reveal that experimental procedures were mainly used in optimizing the performance of DMFC systems. Alternatively, mathematical modelling can be a promising way of finding the best operating conditions for improving the performance of DMFC systems because it is a cost effective approach and also requires less time for its implementation. Therefore, the present work proposes two artificial intelligence methods (Gene Expression Programming (GEP) and M5 model tree) to study the current characteristic of fuel cell with respect to five input operating conditions of DMFC. Performance of the two methods is evaluated against the actual data based on the three statistical metrics, hypothesis tests and cross-validation procedure. The hidden relationships between the fuel cell current and the five operating conditions are extracted by performing a parametric analysis on the model. It is found that the cell temperature has the highest impact on the current characteristic of DMFCs. The insightful information extracted from the model analysis could be useful for improving the working efficiency of the fuel cell.