Research Of Vinyl Acetate Synthesis Reaction Optimization

Vinyl acetate is a colorless, transparent, strong pungent odor liquid. It is currentlyworld’s largest output of50kinds of chemical products. In recent years, due to theincreasing demand, how to further improve the production process and reduce energyconsumption has become a modern enterprise technology innovation content.At first, paper made an optimized plan for the whole design according to the actualcondition; also it analyzed the online real-time monitoring for the synthetic reaction is theprerequisite optimization scheme. Due to the lack of necessary testing equipment, paperdecides to establish a mathematical model to simulate the course of the reaction. Aftercomparison of several commonly used modeling algorithms, article selected neuralnetwork modeling the synthesis reaction. But due to the randomness of the initial weightsand thresholds, leading to the neural network modeling results are unsatisfactory.Against this background，paper through a combination of genetic algorithms andneural network to build the new model. The genetic algorithm is introduced to takeadvantage of its good global search capability to reduce the risk of limited local optimalsolution. At the same time, according to the characteristics of the neural network algorithmto avoid that training is too slow, resulting in not conducive to practical application.Specifically, the first complete genetic iteration of the genetic algorithm to determine agood initial weights and threshold, then conduct the traditional neural network trainingsteps to train the network. By comparing the simulation results，proved that the new modelhas higher fitting precision and prediction performance, accurate approximation andgeneralization ability, while significantly reducing the running time of the model, is moreconducive to the practical application of engineering.For controlled variable adjustment, article first identified which is needed to beadjusted, that is synthesis reactor temperature setpoint. Then based on the previouslycreated synthesis reaction model, article use genetic algorithm to solve the reactortemperature optimization problem, and gives the genetic iteration results and comparisonof the data to prove it.