A laboratory-scale dual-reflux pressure swing adsorption process unit comprised of two beds with intermediate feed position was built and tested for concentrating dilute gaseous feed of methane in nitrogen using Norit RB1 activated carbon. Experimental investigations revealed that the change in blowdown or pressurization step time, feed or purge step time, and heavy product flow rate did not have a significant effect on process performance (in terms of methane purity). However, the process performance was found to be quite sensitive to changes in light reflux flow rate. These experimental results were used to validate the predictions of a mathematical model developed in the frame of the commercial software Aspen Adsim, thus providing an effective tool to elucidate process behavior.