Abstract: To maintain optimum performance of the electrical
response of a fuel cell, a real time identification of the
malfunction situations is required. Critical fuel cell states depend,
among others, on the variable demand of electric load and are
directly related to the membrane hydration level. The real time
perception of relevant states in the PEM fuel cell states space, is
still a challenge for the PEM fuel cell control systems. Current
work presents the design and implementation of a methodology
based upon fuzzy decision techniques that allows real time
characterization of the dehydration and flooding states of a PEM
fuel cell. Real time state estimation is accomplished through a
perturbation-perception process on the PEM fuel cell and further
on voltage oscillation analysis. The real time implementation of
the perturbation-perception algorithm to detect PEM fuel cell
critical states is a novelty and a step forwards the control of the
PEM fuel cell to reach and maintain optimal performance.