Simulação dinâmica de uma mistura de gases combustíveis industriais utilizando controle multivariável sem restrições

Abstract

This paper deals with a case of literature and involves simulation, where currents of combustible gases entering a header, a network of fixed volume of products must be controlled within specified limits. There are two streams which have no fixed composition, TG1 (Tail Gas 1), TG2 (Tail Gas 2), which is commonly-laughing burned so has no commercial value and four fixed composition of feed streams: natural gas (Gas Natural), GR (reformed gas), H2 (hydrogen) and N2 (nitrogen). In the header of these six currents are mixed so that in the left-it will be obtained a gas stream with fuel properties as HHV (Higher Heating Value) and WI (Wobbe Index); according to what would be required by an industry. These properties provide information about the quality of fuel gas, whether he has good performance in burning because they involve the calorific value in their equations. The simulation is done in two software: the chemical process simulator Aspen HYSYS® V7.3 and MATLAB® R2014a with mpctool. To control the flow of the six chains, we used a multivariable controller, MPC (Model Predictive Control), requiring the process model, which was introduced the transfer functions for HHV and WI properties of six chains. The results show that the use of a predictive control can increase the value of property HHV (Higher Heating Value) of 15.73 MJ / Nm³ to 16.75 MJ / Nm³, this indicates that the calorific value of the gas, taking it well better quality.