Study Of De-heavy Tower Control System Based On Embedded System

During the production of crude benzol refining, crude benzene need to be subjected to de-heavy Tower for effective separation of heavy light benzene and benzene, hence the actual production environment of the tower temperature control system requires high precisions. The de-heavy tower has some characteristic include nonlinear, long delay and unpredictable, currently in most of the cases, it controlled by high cost, bulky and less responsive conventional PLC which is using PID temperature control algorithm. The rise time and overshoot of these two performance indicators can't be optimized simultaneously. Therefore, the paper uses multi population genetic algorithm tuning PID parameters to achieve the optimization of the control strategy of multi performance index.Firstly, creates mathematical model by actual production data and builds the simulation system on Simulink. Determines the PID control parameters by production data, Then, the three parameter values were used as the central point to delineate the neighborhood section as the gene pool of genetic algorithm and choosing the appropriate sectional function as the fitness function to guide the evolution direction of the two sub populations. Operation of the two sub-populations were involved in the evolution and tuning PID parameters at different stages of temperature changes, so that the temperature change of the rise time and overshoot both get a good performance optimization.Secondly, Through the introduction of the Smith predictor compensator to avoid the dynamic response delay which is coused by tower's pure delay.Finally, Designed an experimental control system. The system uses dual ARM controller with data processing and algorithm operations separating design scheme to improve the accuracy and real-time of temperature control process. Through the analysis of experimental data of two simulation platform, the conclusion is that the control strategy of the multi population genetic algorithm tuning PID parameters is applied to the process of the tower temperature control with excellent stability and dynamic characteristics. It has broad application prospects.