The Phu Kham copper-gold deposit in Laos is an extremely heterogeneous orebody. The deposit has complex and variable mineralogical, geological and geotechnical properties, which affect plant throughput and metallurgical performance. An increased proportion of highly competent ores expected in future operations at Phu Kham has the potential to limit plant throughput.To evaluate how to maintain current throughput rates over the life-of-mine (LOM), Phu Bia Mining Limited conducted a throughput forecasting and optimisation project for Phu Kham with the assistance of Metso Process Technology and Innovation (PTI). The project involved a review of the current blasting, crushing and grinding processes, and development of a throughput prediction model based on geometallurgical modelling for long-term planning. The scope also included identifying opportunities for increasing throughput and improving overall comminution circuit performance when treating the most competent ore types. A shorter term objective was to identify if and when secondary crushing or other process changes will be required to maintain the target throughput over the LOM.The throughput prediction model consists of integrated site-specific models of the drill and blasting, crushing and milling operations. These models were developed, calibrated and validated using SmartTag ore tracking technology which links the ore source (and properties) in the mine with blasting and plant performance (throughput, grade, recovery, etc) in real-time. Geometallurgical ore domains were defined and blasting and comminution processes modelled using measured plant results, thus enabling short to long-term strategic planning and optimisation.Metso PTI is also implementing a GeoMetso system at Phu Kham, which uses the SmartTag ore tracking technology to continuously collect plant data and automatically update the predictive models and block model in real time. This eliminates the need for further expensive ore characterisation tests, and improves the accuracy and predictive abilities of the geometallurgical models that were developed. This can improve long-term mine planning, and capital equipment purchases can be predicted well in advance of their requirement. In the short-term, the processing plant receives advance notice of the ore type(s) about to be processed and adjustments can be made to operating conditions to optimise plant performance.CITATION:Bennett, D, Tordoir, A, Walker, P, La Rosa, D, Valery, W and Duffy, K, 2014. Throughput forecasting and optimisation at the Phu Kham Copper-Gold Operation, in Proceedings 12th AusIMM Mill Operators' Conference 2014 , pp 381-390 (The Australasian Institute of Mining and Metallurgy: Melbourne).