Background At the Thomas Playford Generating Station of the Electricity Trust of South Australia (ETSA) at Port Augusta, difficulty is experienced in maintaining maximum rated output for extended periods. Observations by ETSA staff and research conducted by CSIRO have defined reasons for these difficulties and indicated that the problem might be reduced if the combustion efficiency could be increased. Objective The Australian Mineral Development Laboratories undertook to study the operation of pulverised coal burners of the type used at Port Augusta and to suggest ways of increasing the combustion efficiency. The work involved an investigation of the operating characteristics of the burners and the construction of a cold model of one of the burners with which to study flow patterns and to test modifications. Following upon the investigation of the model, modifications were to have been made at the power station. Summary of Work Done Experiments conducted on No. 1 boiler of the Port Augusta “B” Station showed that variations recorded in the oxygen level were probably due to variations in coal quality or quantity or stratification of the gas stream, rather than to variation in the air supply or a fault in the analyser. Analyses of the fuel supply indicated that crossflow between the two ends of the duplex pulveriser mills was not sufficiently serious as to warrant modifications to prevent it. The distribution of solids in the ducts leading to the burners was found to be asymmetrical. The tests revealed that there was no simple correlation between mill age and performance. Work on the model burner showed the flow pattern probably existing in the burner tubes and allowed development of methods of improving the uniformity of solids distribution in the fuel air mixture issuing from the burner mouth. Conclusions The performance of pulveriser mills is not as dependant on time in service since overhaul as had been though, and measurements have enabled the period between overhauls to be increased. The distribution of solids in the gas stream leaving the burners was found to be very uneven. In the model the solids, which should be spread around the full circumference of the tube, a distance of 61 inches, were confined largely to an arc of about 6 inches in length. This condition is unfavourable for combustion as air penetration to the centre of the band of coal will take some time, during which the particles will pass out of the hottest part of the flame. This could be at least partially responsible for the unburnt carbon found in the ash box and the precipitator dust. Extension of the band of solids as far as the furnace wall could lead to hot spots and the associated problems of corrosion and slag deposition. Suggested modifications to burner design have not yet been adopted at the power station. Installation of narrow longitudinal baffles in the burner tube gave sufficient improvement in the cold model to warrant trial on at least one boiler in the “B” station. Recommendations It is recommended that a full study be made of the performance of one or more pulveriser mills over an extended period. The study should include product particle size analyses, assessment of coal hardness and maintenance of records of mill running time and feeder speeds. This could lead to a further useful increase in life of a mill between overhauls. It is also recommended that modifications be made to the design of the burners in the Port Augusta “B” station. Some of the modifications tested on the model offer a considerable improvement over the existing design, and the use of multiple longitudinal baffles in the primary burner tube is recommended as a logical first step.