The Study On Optimal Grade Transition Of Continuous Polymerization Process

In industrial production, grade transition occurs frequently as the reactor is usually required to be operated under different conditions for producing polymers of different grades, which leads to the reduction of normal production time and generation of large amounts of off-specification polymers, as a result, the economic performance of enterprise will be affected. Therefore, optimal grade transition is important to polymer production. In this paper, the author focuses on styrene polymerization and studies the optimal grade transition of continuous polymerization process.Firstly, an improved multi-objective particle swarm optimization algorithm is proposed in consideration of the shortcomings of traditional multi-objective optimization algorithm. The Kent mapping is adopted to initialize the population, meanwhile, crowding distance is used to maintain the size of external archive set, and the number of Pareto optimal solutions can be controlled. Small probability mutation mechanism is added to strengthen the global searching ability; the particle with largest crowding distance from the external archive is chosen as the global extremum to guarantee the multi-directional evolution of particle swarm. This algorithm is proved to be effective, which has ever been applied in multi-objective optimization of knapsack.Secondly, the author builds a mechanism model of polymerization process on the basis of polymerization mechanism of reaction and polymerization kinetics from the perspective of mass conservation and energy conservation. The Mn, Mw and PDI of the polymer molecular is calculated according to the moments of inactive polymer and active polymer.Finally, the author takes the grade transition of styrene polymerization as an example to study on optimization strategies of sectional grade transition in continuous polymerization by virtue of particle swarm optimization with single as well as multiple objectives. The verification result of both single input case and multiple input case shows that the optimal transition scheme concluded through optimization of the single and multi-objective particle swarm can remarkably reduce the transition time and consumption of raw materials compared with step transition.Compared with single objective optimization method, multi-objective optimization provides a set of optimal solutions. Therefore, enterprises can take into consideration both transition time and tradeoffs in choosing the best solution to ensure optimal grade transition which means the remarkable reduction of transition time and tradeoffs and will increase normal production time, enhance production efficiency and increase corporate profits.