摘要 |
PURPOSE: A method for forecasting a width of a main flap using a neural network is provided to forecast the width of a main flap with accuracy by inputting information as the kind of a steel, a furnace temperature, a width of a mold, the second cooling pattern and a casting velocity through a neural circuit network. CONSTITUTION: A kind of a steel, a furnace temperature, a width of a mold, the second cooling pattern and casting velocity are collected and detected by communicating with a temperature detecting unit(7) detecting the furnace temperature, a mold width controller(8), a driving roll controller(9), a main flap width measuring device(10), a programmable logic controller(PLC)(11) and a process computer(12). Data of the main flap width actually produced in a casting process and a drawing process in accordance with the above conditions are collected. A neural network(13) constructed based on the information collected in the above stages is communicated with the process computer(12) and stores a variation of the main flap width of the steel corresponded to the various conditions. The equipment is controlled for producing the optimum main flap width respond to the main flap width inputted through the most approximate value.
|