OneSteel Whyalla's grate-kiln-cooler pelletising plant was commissioned in 1968 processing haematite feed from the local South Middleback Ranges mining area. The existence of a magnetite resource beneath the haematite ore at Iron Duke had been known for some time. In 2004, a decision was made to covert the primary iron source for the Whyalla steelworks from haematite to magnetite - known as project magnet. The intent of the project was to convert the feed to the steelworks from haematite to magnetite and to free up the remaining haematite resource for export. The pelletising plant was cut over to a magnetite feed in 2007.The thermochemical differences of magnetite compared to haematite are threefold:1. the introduction of an exothermic reaction during conversion of magnetite to haematite,2. the absence of bound water which existed within the haematite, and3. lower gangue content requiring lower fluxing levels and therefore lower calcination energy demand.The combination of these differences results in a significantly different induration requirement for magnetite as compared to haematite. The change therefore required extensive research and test work to design modifications to the furnace and to ensure appropriate risk mitigation for the Whyalla steelworks.Preliminary pot-grate test work was performed to define the pellet chemistry and confirm the suitability of the magnetite to produce pellets suitable for the OneSteel blast furnace. A heating curve was developed by use of computer-based heat and mass balance modelling. Extensive pilot scale grate-kiln-cooler test work was then performed to confirm the heating curve and to verify and fine-tune the pellet properties. Computation fluid dynamics was used to design the physical changes to the furnace required to achieve the proposed heating curve.Following installation of an additional preheat zone, including a wall and dilution air fans, into the preheating furnace, and installation of an additional wall into the cooler, the modified furnace was commissioning in December 2007. The plant performed as expected during commissioning with the actual heating curve matching the proposed curve and the produced pellets within blast furnace specifications.