Otimização e controle preditivo na distribuição de combustíveis siderúrgicos

Abstract

In an integrated steel mill, the gases, byproducts of steel production processes can be recovered as fuel in the plant itself, for example, for use in industrial furnaces, steam generation in boilers and electric power generation in thermal power plants. For efficient use of these fuels, different classes of problems need to be solved at different levels of process control hierarchy. In the optimization level, it is necessary to determine solutions for the efficient distribution of byproduct fuels, which have different characteristics between the production and consumption profiles, with various types of consumers, limited storage capacity (in gas holders) and restrictions on their use in thermoelectric plants. The level of advanced control must be able to handle the hybrid dynamic behavior of the thermal power plant boilers (determined by instruments and control elements with discrete and continuous characteristics, subject to a series of logical rules for operation).This thesis deals with the development of a solution to the problem of byproduct fuel distribution in integrated steel mills, which includes a mixed integer linear programming model (MILP) for maximizing energy efficiency of the process and a model based predictive control (MPC) strategy applied to hybrid systems, for integration of the optimization level with the regulatory control level. Original contributions are carried out, among which stand out: (1st) improvements in MILP model compared to previous approaches; (2nd) development of a heuristic procedure to determine appropriate values for penalties applied in the objective function of the MILP model; (3rd) modeling of a burner management system (BMS) of an industrial boiler by means of the theory of hybrid systems; (4th) integration between the optimization level and the advanced control level to solve the problem of fuel distribution system at integrated steel plants.