Desenvolvimento de um modelo de incrustação e aplicação de algoritmo genético na programação de limpezas de tanques de resfriamento.

Abstract

In the process of caustic soda production, one of the stages is the cooling of the sodium hydroxide solution. The cooling of the sodium hydroxide solution is made in a series of tanks that use frozen water and water of cooling tower to reduce the temperature of the solution until the specified value. Each tank is endowed with agitator and coil of cooling. The cooling water flows in the interior of the coils in countercurrent. The water of cooling tower is used in the first tanks, where as the frozen in the last tanks. One of the great problems of the industrial processes is the fouling formed in the equipment. In the system of caustic soda water cooling of the BRASKEM, fouling if they form due to crystallization of leave around the coil diminishing the global coefficient transference heat. An asymptotic model in function of the time for fouling in the tanks it was adjusted to determine optimum moment where the tank must be clean. To optimize the cleaning of the tanks in relation to the period of time and the choice of the tank is and to minimize the number of periodic cleanings they are the objectives of this work. The objective function is calculated by the program based on a model for the simulation of this system of cooling with developed previously for this project and integrated model of asymptotic fouling function to other subprograms developed in MATLAB that use the genetic algorithms to choose the best solutions for the system. From the genetic algorithms some excellent places had been determined analyzing the values of the temperature of the solution in the last tank and the number of cleaning per day, as well as effect in the thermal load and the difference of temperature enters the water of cooling and the solution. With the use of the genetic algorithms it is being possible to find optimum cleaning schedule for the tanks.