Parameter estimation of proton exchange membrane fuel cell system using sliding mode observer

Parameter estimation for health monitoring and sensorless scenario is achieved online via model-based robust state observer. The proposed dynamic algorithm observes a state (output voltage) of fuel cell system (FCS) in the presence of uncertainties and disturbances. Using this observation, membrane conductivity is estimated. The conductivity is function of water content and temperature which gives the value of water content analytically. The water content represents two important faulty modes, flooding and drying in proton exchange membrane (PEM) FCS. The water content can generally be measured through humidity sensors or other techniques. In the case of sensors, their size and cost restrict in-situ measurements whereas regarding other techniques, the extractive sampling makes measurement process slow and intrusive. Moreover, the existing measurement techniques have the issue of accuracy which is of prime importance in control and diagnostics spectra. The sliding mode technique is employed for the design of observer. The technique requires a dynamic voltage model that is developed through extensive mathematical modus operandi. The computer simulation confirms that the estimates are quite similar to nominal value. Experimental extracted range of parameter verifies the magnitude of estimated parameter and its precision is validated through offline calculation of the parameter using a model available in the literature. The observer can replace the humidity sensor which results in ridding of expensive and hard measuring instrumentation. The water content parameter estimation can provide a foundation for design of fault diagnostic schemes in PEMFCS.