A sustainable coal industry needs to achieve a number of objectives including improvements in utilization of deposits, energy efficiency, and environmental performance during combustion. The "Clean Coal Concept" promises to combine the secure and low cost electricity supply offered by hard coal and lignite with adequate environmental precautions. In this context, modern quality control systems, which are greatly based on the on-line analyzers, play an important role in achieving sustainability targets. This paper examines the possibility of using commercial on-line analysis systems for monitoring the mineral matter content of low quality lignite mined from the lignite basin of Megalopolis, Peloponnese, Greece. The effect of the accuracy of the on-line analyzers to process control, when used for the characterization of low rank coals with complex and variable composition, is investigated by carrying out numerous bench- and pilot-scale trials. Pilot-scale trials were based on a dual energy gamma-ray transmission analyzer, which was installed on the conveyor belt that transports lignite from the mine pit to the homogenization stockyard. All measurement data were compared to data gathered during the realisation of similar trials in the lignite mines of Ptolemais Basin, Northern Greece. Results indicated that the accuracy of the on-line measurements was not satisfactory and did not allow lignite quality monitoring in real time. The achieved inferior accuracy of the on-line measurement's accuracy, compared to previous applications at other mining sites, was related to the intense variation of the mineral matter content of lignite and lignite composition, which distorted the calibration of the analyzer. The latter is based on certain assumptions regarding the average mass absorption coefficient of the organic and mineral matter contained in the lignite. Further experimental work is needed to investigate solutions for successful implementation of this method to low grade lignites that exhibit large variation in mineral matter content and composition. (C) 2009 Elsevier B.V. All rights reserved.