1 – September 2012 – Steel Times International Reheat furnace www.steeltimesint.com A three-dimensional Computational Fluid Dynamic (CFD) modelling approach developed by Centro Sviluppo Materiali together with Tenova to develop the FlexyTech® range of low NOx emission burners has been extended to the application of a complete reheating furnace. By A DellaRocca* M Fantuzzi*, V Battaglia** & E Malfa** Advanced design methods for rotary hearth furnaces *Tenova LOI Italimpianti, Italy, **Centro Sviluppo Materiali SpA, Italy ENVIRONMENTAL concerns and rising prices of fossil fuels have focused the steel industry’s attention on minimising emissions and improving heating efficiency in reheat fur- naces. Additionally with the rise in competitive- ness within the world steel market product quality can be a key strategic factor to gain a market, a factor which arises in the reheating furnace by ensuring the correct heating rates, ultimate temperature, uniformity of tempera- ture and furnace atmosphere. All those targets need to be fulfilled without increasing capital expenditure and operating costs beyond acceptable limits. As a conse- quence, a successful project for a reheating fur- nace must accomplish all these targets to min- imise energy consumption, limit emissions and ensure product quality as far as is practically possible. This scenario poses various questions when designing or modifying a reheating furnace, many of which can provide valuable answers through integration of advanced design and verification methods into the common engi- neering workflow. Various attempts to evaluate the heating process inside a furnace have been made in past years with different levels of detail and accura- cy. As it is well understood that radiation is the main mechanism of heat transfer involved in heating the charge, simplified models are com- monly used to analyse performance and effi- ciency of reheating furnaces. The well-estab- lished zonal methods range from single or two dimensional models for preliminary estimation of charge temperature to three dimensional models able to account for edge effects in slab and square bloom or billet passing through a walking beam furnace as well as the cooling effect of the skids that carry the charge through the furnace [1-4] . Nevertheless, such models cannot give an indication about the local combustion process- es which develop inside the furnace, nor can they give an indication of expected emissions levels. From the experimental side, existing methods of measuring cannot provide continu- ous measurements of temperature inside the furnace. Trailing thermocouples or data loggers can be used only with test charges to give an indication of the heating process at a few points measured inside the charge, while optical pyrometers or thermal cameras give informa- tion only about the surface temperature of the charge and their accuracy is greatly affected by the presence of the surface oxide layer (scale) on the charge. Therefore in recent years two main developments are on-going thanks to con- tinuously increasing computational power: These are: Fig 1 TRGX flame temperature iso-sur- face coloured in terms of CO 2 mole fraction. Top con- ventional long flame burner bottom short compact burner Fig 2 FlexyTech TRX- 4 burner Ɂ Online three dimensional control tools based on existing zonal models extended to simulate transient radiation heat transfer [5] . Ɂ Three dimensional Computational Fluid Dynamic (CFD) simulations of the com- plete furnace, coupling the reacting flow combustion models developed for single burners, the heat radiation model and conju- gated heat transfer models of charge heat- ing [6,7] . This article is focused on the definition of a CFD model able to simulate the large rotary hearth furnace at the TenarisDalmine pipe mill and thus able to evaluate the fluid dynamics of the furnace gases, quantify the thermal and chemical species inside the furnace and provide a fine scale representation of the heating process of the charge. This rotary hearth is equipped with TRGX flameless regenerative burners and TRX flame- less roof burners with the aim of minimising polluting emissions and maximizing furnace efficiency through pre-heating the combustion air to a high temperature. The established CFD methodology is inte- grated with the computational fluid dynamics approach that Tenova currently applies for the design, verification and continuous improve- ment of combustion systems. Using this successful combination combined with different levels of analysis, a complete multiscale methodology for furnace design is presented and the excellent results on an indus- trial installation are presented. CFD methodology The physical and numerical model currently adopted by Tenova for verification of burner design, and continuous improvement [7, 8] have been extensively validated against experimental data from CSM test rigs. A complete database of CFD simulations for all the range of Tenova’s ‘FlexyTech’ burners of various sizes and at discrete turndown levels is continuously populated with newer data from the CFD activ- ities undergoing in the R&D projects between CSM and Tenova. This creates the fundamental knowledge base that is required for the success- ful application of different combustion tech- nologies such as flameless and regenerative combustion. In particular, regenerative burners pose tech- nological issues for rotary hearth furnaces and also generally for any reheating furnace with narrow chambers. To have the characteristic cycling behaviour between the regeneration and firing phase for each burner pair, it is nec- essary to install burners with double the design output size [16] as one of each burner pair is inoperative during the regeneration cycle. As a consequence, the flame length increases and serious questions arise about the possible ingestion of partially reacted fuel/air mixture by the burner on the opposite wall, with result- ing detrimental effects on both combustion efficiency and regenerative bed maintenance. This effects is of particular concern in relative- ly narrow chamber furnaces such as those of a rotary hearth furnace. To evaluate this effect and design flameless regenerative burners able to cope with these