When dealing with molten metal anywhere from 800-1500 ͦC (1500-2700 ͦF) a lot can go wrong. Splashing can cause spills that endanger workers and equipment. Meanwhile, improper water contact at the casting facility can create steam, or hydrogen explosions in some extreme cases.
To ensure the safety of operators and equipment, however, experimentation is not an option. Testing these conditions would damage the temperature gauges and the risk to experimenters is too great.
For companies such as TBR, conditions in various locations of the molten metal were often unknown and past designs were mainly governed by a trial-and-error approach. TBR decided, therefore, that simulations must be used to ensure that all goes well in the casting house.
The system is at greatest risk to splashing when liquid metal exits the furnace from a tap-hole into runners, which separate the slag from the metal. This risk is mitigated using a concrete and cast iron roof over the runner. The concrete acts as an insulator protecting the iron from the molten metal.
Runner cross section. With inner concrete layer in contact with molten metal and outer layer providing support.
However, these runner roofs do not last long in these extreme conditions, usually about a month. As replacing them can be expensive, trial and error design methods are impractical from a cost-return perspective.
Simulations of the runners started by mimicking the real life conditions in a pre-heat. The simulated runners were pre-heated to 500 ͦC (930 ͦF) using gas burners. The layer of concrete, which would be in contact with the molten metal, was dominated by conductive heat transfer, and air convection was ignored during the pre-heat.
The model then included the contact with the molten metal for 75 minutes. Heat transfer was set by conduction and convection of the air in the system. Simulation showed that it took 300 seconds for the contact walls to reach steady state temperature with the liquid metal. After this point, conduction was slow between the concrete layers. Additionally, it was found that after 500 seconds the air was still allowing for future simplification of the model by ignoring air convection.
Temperature profile of the runner with liquid metal contact shows little change between 300 and 4500 seconds.
The simulation was then extended to a 7-day long heating and cooling cycle. The cycle consisted of 75 minutes of metal contact followed by 75 minutes without molten metal. It was found that the outer metal shell of the runner never reached over 80 ͦC (175 ͦF) and the air in the open space never reached over 400 ͦC (750 ͦF). This temperature was then verified by testing the outer temperature of an actual runner using thermal imaging.
The results showed that TBR were over engineering their runners. They assumed the air temperature in the cover was much higher than it was. They were therefore able to reduce the thickness of the concrete within their runners, saving money and learning a little more about the processes they perform on a day-to-day basis.
Cover image courtesy of Bigstock
Images and Source courtesy of COMSOL