Research On Convergence Performance For Process Simulation Of Stream Circulation System

There widely exist stream circulations in chemical process system.It is difficult for sequential module method widely used in process simulation to deal with stream circulations.Stability theory in control theory and pulse elimination theory for singular system is used to solve the problem.The research of the convergence performance for stream circulation system is of theoretical value and practical significance.Firstly,the mechanism models for each unit are established,such as the riser and regenerator of FCCU(Fluid Catalytic Cracking Unit),the reactor and distillation tower of toluene oxidation unit.Then the process simulation was studied respectively.Secondly,a convergence study for the steady-state process simulation is carried out for stream circulation system.According to the form of mathematical model,chemical plants are identified as forward model and reverse model,and the variables in stream circulation are defined as iterative variables and convergence variables respectively.The stream circulation is broken at the convergence variables,and the convergence variables are calculated from the iterative variables by the forward model and the reverse model respectively;the convergence errors are used for the correction of iterative variables through the convergence gain coefficients,and then the iterative equation is obtained.The convergence gain coefficients of the iterative equation are determined using the control stability theory;the iterative equation is linearized and the Routh criterion is used to determine the stable range of the convergence gain coefficients;when the convergence gain coefficients is within the stable range,the iterative equation must converge.Finally,the dynamic simulation of stream circulation system is studied based on the pulse elimination theory for singular system.For the impulse characteristics brought by the stream circulation in the dynamic simulation,a virtual circulation loop is added between the forward/reverse model to eliminate the impulse by using the output feedback.The feedback convergence gain coefficients in the virtual loop are determined using the pulse elimination theory for singular system.The system is discretized to obtain the dynamic iteration equation.When the convergence gain coefficients satisfy the impulse elimination condition,the dynamic simulation will not show pulses when the process conditions are changed and the dynamic iteration finally converges to the new steady-state value.