Simulation of an Industrial Furnace for Aluminum Annealing

be28adfff47893c4519c1307dc6b8866_L

SimTec in cooperation with EL.K.E.ME. (www.elkeme.gr), the research center of VIOHALCO Group of companies, constructed a mathematical model for the simulation of batch annealing furnaces for aluminum strip in coiled form installed at the EL.V.AL. production company. The coils are formed by winding aluminum strips, of various thicknesses and widths after a cold rolling process. Annealing may last up to 24 [h], depending on the coil size and strip thickness. These furnaces have dimensions 7.0×5.5×4.5 [m] and the thermal energy consumed is approximately 1.2 [MW].

Annealing is crucial for the improvement and uniformity of the mechanical properties of the metal (grey cylinders in 1st figure) following the straining of cold rolling and is performed in nitrogen atmosphere to avoid oxidation due to the elevated temperature. The heat to nitrogen is provided by 12 direct burners which throw the flue gases inside perpendicular U fire tubes (red in 1st figure). In order to prevent surface overheating of the coils, an internal Π-shaped curtain traverses the whole length of the furnace (yellow in 1st figure). Nitrogen flow is achieved by the action of two horizontally-arranged axial fans (blue in 1st figure), so that a high heat transfer coefficient due to forced convection exists between nitrogen and aluminum. Nitrogen and aluminum (surface and internal) temperature is monitored by thermocouples, and burner and fan operation is controlled by an automation system, so that the desired annealing temperature curve (nitrogen temperature vs annealing time) is realized.

The study included:

  • Construction of geometry from available 2D mechanical drawings and mesh generation.
  • Construction of the mathematical model in FLUENT CFD package for the solution of turbulent nitrogen flow and heat transfer in nitrogen and solids including radiation. All material properties used were taken functions of temperature.
  • Thermal conductivity of the coils was not uniform but orthotropic, due to the existence of the gap between successive strips of aluminum.
  • Fan operation was realized by importing into FLUENT’s special fan model their characteristic curves (ΔP vs Q data points).
  • The automatic operation of fans and burners according to the programmed annealing curve was achieved by the authoring of external routines in FLUENT, so that the actual process is faithfully simulated.

The initial target of the study and the sensitivity analysis of the various process operating parameters (strip dimensions, number of coils, alloy type, fan or/and burner failures), was to map the performance of the annealing furnace, whereas a subsequent goal is to optimize the process (achieve temperature uniformity and annealing time reduction).

furnace2