Simulations of a Furnace Brazing Process Using Tera-Scale Computing

摘要

Thermal responses in a production brazing furnace can be very difficult topredict or measure due to changing loading and heating conditions. An efficient approach to determining acceptable furnace brazing cycles is to apply computational models that can solve these potentially complex thermal problems. Recent developments in massively parallel computing have facilitated finer-scale temporal and spatial thermal analyses. Sandia National Laboratories has conducted terascale simulations to thermally characterize a production brazing process. The model accounts for thermal radiation and conduction as the principal heat transfer drivers in the batch-style hydrogen furnace. Computed peak temperatures are within one to two percent of the programmed and measured values. Thermal responses in the work zone are particularly sensitive to the thermal enclosure, defined by the furnace's heating elements. A two percent change in the radiation enclosure can yield up to an eight percent change in the peak brazing temperature. Computational data are less sensitive to differences in the inputted materials thermal properties than changes to the thermal enclosure. Examples of how the thermal model can be used to determine optimal loading and heating conditions for production brazing are discussed.