基于神经网络模型的对二甲苯氧化过程优化

Model and optimization of para-xylene oxidation process based on neural network

根据PTA工厂生产操作数据,用人工神经网络的反向传播算法建立了氧化过程尾氧和尾气二氧化碳生成速率的预测模型。在此基础上,利用逐步二次规划(SQP)法对预测模型进行了寻优,获得了优化的工艺条件。工业应用表明,它实现了氧化产品质量稳定条件下最低的燃烧消耗量。

Based on the real data from the purffed-terephthalic acid （PTA） manufacturing process, an artificial neural network model was developed for predicting the flow rate of off-gas oxygen and carbon dioxide in the para-xylene oxidation process. The operation condition of para-xylene oxidation was optimized by sequential quadratic programming. The commercial production showed that it could not only satisfy the requirement for PTA purity, but also reduce consumption of para-xylene and acetic acid combustion.