【摘要】 针对中药生产提取工段控制量多、耦合强的特点,以减小提取罐中药液温度稳定时间、提高挥发油的回收量、减小能源消耗总量和增加有效药液提取量4个指标作为优化目标,提出一种利用神经网络对控制器进行多目标优化的设计方案,神经网络的权值和结构分别用遗传算法和变结构算法训练。实践结果表明,设计出的中药生产提取工段的控制器具有良好的综合性能:缩短了药液温度达到稳定值的时间,提高了药液和挥发油的产量,降低了总能耗,很好地完成了多项目标的优化。更多还原

【Abstract】 A multi-objective optimal control scheme,which introduced a neural network to optimize the controller,was put forward.Meanwhile,genetic algorithm and variable structure algorithm were used to train the weight parameters and structure of neural network.Because of excessive controlling variables and strong coupling in withdrawing Chinese traditional medicine,debasing the steady time of the temperature in the withdrawing pot,increasing the output of the naphtha,decreasing the energy consumption and enhancing the amount of effective liquid in withdrawing were taken as four optimal objectives.The results show the designed controller of withdrawing Chinese traditional medicine has better synthetic performance and the optimizing of objectives,such as less steady time,lower energy consumption,more naphtha and effective liquid are completely achieved.更多还原