热偶精馏过程模拟与优化的改进

热偶精馏是一种新的节能方式，但传统的热偶精馏模拟计算过程繁复，且基于传统模拟过程的优化方法也难以得到良好的可行解。针对这一问题，本文利用人工神经网络和遗传算法的特点，提出将二者结合以改进热偶精馏过程的优化方法，并将该方法应用于丁二烯分离及乙腈回收流程的研究中，先由神经网络建立黑箱数学模型，再由遗传算法优化热偶精馏的操作参数，通过计算结果比较，表明该方法可以迅速得到优化变量和目标函数，而且获得全局最优解。

The advancement of simulation and optimization for thermally coupled distillation

The special advantage of the thermally coupled arrangement is the achievement of the energy savings. Despite their advantages, designers have been reluctant to use thermally coupled columns. This reluctance can be attributed to fear of control problems and to the lack of design procedures. In this paper, a new approach using genetic algorithm and neural network for the optimization of the thermally coupled distillation is presented. A practical example of the process of butadiene through extractive distillation using acetoni-trile as solvent is used to demonstrate the algorithm. With the proposed model and optimization arithmetic, mathematical model can be calculated, decision variables and target value can be reached automatically and quickly.