论文部分内容阅读
在15~30MPa和303~323K条件下,用超临界CO2流体萃取沙棘籽油。结果表明,最高沙棘油收率(30MPa,308K)可达到90%以上。对过程进行动力学模拟,建立了超临界萃取过程的人工神经网络(ANN)模型。以MATLAB软件为平台,编制了SFE-ANN模拟程序系统。采用3层BP网络结构,以压力、温度、萃取时间为输入,以萃取出油量为输出对网络进行训练。由此得到的网络可以对萃取速率和单位时间床高方向的萃取出油量进行准确的模拟和预测。与实验结果比较证明,训练样本集误差为0.2%,测试样本集误差为0.5%,模拟误差小于6%。
Seabuckthorn seed oil was extracted with supercritical CO2 fluid at 15 ~ 30MPa and 303 ~ 323K. The results showed that the highest seabuckthorn oil yield (30MPa, 308K) can reach more than 90%. The process of dynamic simulation, the establishment of the supercritical extraction process artificial neural network (ANN) model. With MATLAB software as a platform, the system of simulation program of SFE-ANN has been compiled. Using 3-layer BP network structure, with pressure, temperature, extraction time as input, extract the output of oil as output network training. The resulting network allows for accurate simulation and prediction of extraction rates and bedside extraction oil volumes per unit time. Compared with the experimental results, the error of the training sample set is 0.2%, the error of the test sample set is 0.5% and the simulation error is less than 6%.