Resumen

The complex world of polymerization has become an important topic for research due mainly to the great investments it demands and the difficulty in operation of the huge plants that exist to produce polymers. Many papers have been written about this, most of them focusing on the reactor performance and its stability. In the present work a model and simulation for the high pressure polyethylene tubular reactor has been made, but taking into account not only the performance of the reactor by separate but also its interaction with the whole mass balance, in such a way that an integrated model is obtained to study various operating conditions and their effect on exit monomer conversion, peak temperature and number and weight degree of polymerization. At the end, a multivariable control strategy has been implemented (LQR by using Kalman filter) to be compared with PID control (decoupled dynamic PID). It is important to remark the use of MATLAB/SIMULINK software, to model and simulate the complete process. The aim of this study is to make a suitable model for control purposes, in order to try, in the future, different schemes since the point of view of control. The idea is to try new approaches of control strategies like neural model based predictive control which may appear to be promising, due to the complexity of the system. The mass balance involves 13 CSTRs, but more of them can be added to the model. Also, the reactor is simulated using a set of 10 highly nonlinear stiff differential equations, but of course, more equations can be added. In this way, this is a good exercise in the sense that the integrated model (mass balance-reactor), takes into account the complete set of variables intervining in the process.