Fluidized bed combustion technology has been widely used as the new, flexible, multi-fuel boiler for waste combustion and energy recovery from low grade fuels. However, problems such as low thermal efficiency, high emissions, bed agglomeration etc. are still encountered in the operation of fluidized beds. Valuable experiences were gained from two case studies recently conducted regarding wastes combustion in industrial scale fluidized beds. In the first case, the performance of a fluidized bed combustor for energy recovery from oil sludge was evaluated during the commissioning trials. Apart from the sludge characterization and bed material analysis, the combustion efficiency, solid flow balance and on stack emission of CO, SOx and NOx were investigated, as well as the fluidization quality. Although the system was operated with good combustion efficiency (> 99.91 %), sulfur dioxide emission (> 1000 ppm) was found to be substantially higher than the allowable discharge limit. It was recommended to increase the limestone feed rate in order to meet the SO2 emission standard, and subsequently, installation of a cyclone is suggested to remove the potentially significant increase in ash and fine particles. The second case study focused on the bed agglomeration observed in a fluidized bed incinerator where a burning blend of three wastes (i.e. carbon soot, biosludge and fuel oil) is involved. To understand the mechanisms and related chemistry, several analytical approaches are employed to identify the bed materials (fresh sand and degrader sand) and the clinkers formed from full scale incinerator tests. The formation of clinker is believed to follow the mechanism of partial melting and/or reactive liquid sintering. The effects of temperature and blending ratio are tested in a muffle furnace. Carbon soot is believed to be more susceptible than the other two fuels. Thermodynamic multi-phase multi-component equilibrium (TPCE) calculations predict that the main low melting point species are predominant under the oxidizing condition, suggesting that reducing conditions might be favorable to restrain bed agglomeration. This study provides valuable information for better understanding of the chemistry related to clinker formation; it also helps in developing methods for control and possible elimination of the bed agglomeration problem for the design fuels. (C) 2004 Elsevier Ltd. All rights reserved.