超声波天然物提取过程建模、频率优化及应用研究

发布时间：2018-05-10 07:31

本文选题：超声波提取 + 建模　；参考：《江南大学》2017年博士论文

【摘要】：超声波对天然物(本研究主要指中药材和食材等天然植物)的有效成分提取是一种效率高、残留少、无污染的绿色提取技术。近年来,随着超声波天然物提取工业规模的不断扩大,迫切需要对提取过程进行先进的优化和控制,超声波天然物提取过程的模型化是过程优化及控制的基础和前提。目前国内外学者对超声波天然物提取过程的动力学建模做了较深入的研究,然而利用机器学习理论对提取过程进行建模的研究还不够成熟。另外在天然物提取效率方面,在连续宽频带范围内针对超声波天然物提取过程的频率优化问题,尚未有文献报道。因此,根据超声波天然物提取过程的实际情况,建立实用的提取过程模型,同时对天然物提取过程的超声波频率进行优化都具有重要的科学意义和实用价值。鉴于此,论文对超声波天然物提取过程的动力学和软测量建模以及超声波频率优化及其应用等问题进行了较深入地研究,主要工作如下。(1)针对现有的动力学模型未考虑超声波频率对模型的影响,以传质动力学模型和超声波强化机理为基础,通过引入超声波频率,提出一种改进的超声波天然物提取过程的动力学模型。进一步,为验证模型的有效性,通过超声波甘草酸提取实验,在获得最佳提取变量的基础上,分别建立了关于甘草酸浓度与超声功率、超声波频率和提取温度的动力学模型,实验和仿真结果表明,该模型是可行且有效的。(2)针对动力学模型存在的通用性差和可移植性不强问题,以支持向量回归(SVR)理论为基础,建立了基于支持向量回归的超声波甘草酸提取过程的预测模型,预测结果验证了其有效性。进一步,针对支持向量回归模型的训练时间过长等问题,在分析最小二乘支持向量机(LSSVM)理论的基础上,构建了最小二乘支持向量机的超声波甘草酸提取过程的预测模型,并与支持向量回归模型的预测结果进行了对比分析。(3)针对最小二乘支持向量机建模中参数优化问题,通过引入动态步长调节因子和混沌优化算法,提出了一种混沌动态步长果蝇优化算法(CDSFOA),利用马尔科夫收敛性分析理论,证明了该算法收敛于全局最优解,并通过实例仿真验证了算法的有效性;基此,利用该算法对模型参数进行优化,建立了基于CDSFOA-LSSVM的超声波甘草酸提取过程的预测模型,通过与支持向量回归和最小二乘支持向量机模型的预测结果进行对比分析,得到该模型具有更快的训练速度和更高的预测精度。(4)针对上述模型不能在线预测的问题,以无偏置LSSVM和在线LSSVM为理论,提出了在线无偏置LSSVM算法。进一步,针对多目标输出问题,在结合在线无偏置LSSVM算法和多输入多输出LSSVM算法基础上,提出了一种多输入多输出在线学习无偏置LSSVM算法。为进一步提高该算法的运算速度和精度,通过引入加权因子和利用在线递推学习方法,对算法的递推公式进行了改进。基此,为同时预测天然物矛卫中的芦丁和槲皮素两种有效成分的浓度,建立了超声波多目标预测模型并得到验证。(5)针对目前超声波提取过程存在频率单一和无法连续选择的问题,提出一种超声波频率优化方法,即先在宽频带内对最优频带进行大范围的粗搜索,而后在获得的最优窄频带内对最优频率进行小范围的细搜索,通过连续自动地搜索最优超声波频率来实现提高超声波天然物提取效率的目的。基于该频率优化思想开发的超声波提取系统,分别对两种天然物,即食用番茄和中药材槐米的有效成分进行了提取实验和频率优化研究,得到了各自的最优超声波频率并取得了较好的提取效率。
[Abstract]:Ultrasonic extraction of natural materials (natural plants, such as natural plants and natural plants, such as Chinese medicinal materials and materials, etc.) is a green extraction technology with high efficiency, less residue and no pollution. In recent years, with the continuous expansion of the industrial scale of ultrasonic natural extraction, it is urgent to optimize and control the extraction process, and the ultrasonic natural products are urgently needed. The modeling of extraction process is the basis and premise of process optimization and control. At present, scholars at home and abroad have done a lot of research on dynamic modeling of ultrasonic natural extraction process. However, the research on the extraction process by machine learning theory is not mature enough. In addition, in the aspect of natural extraction efficiency, continuous broadband frequency is used. The frequency optimization problem in the range of ultrasonic natural extraction process has not been reported. Therefore, based on the actual conditions of the ultrasonic natural extraction process, a practical extraction process model is set up. At the same time, it has important scientific significance and practical value to optimize the ultrasonic frequency of natural extraction process. This paper studies the dynamics and soft measurement modeling of ultrasonic natural extraction process, the ultrasonic frequency optimization and its application. The main work is as follows. (1) the current dynamic model does not consider the influence of the ultrasonic frequency on the model, which is based on the mass transfer dynamics model and the ultrasonic strengthening mechanism. On the basis of the introduction of ultrasonic frequency, a dynamic model of improved ultrasonic natural extraction process is proposed. Further, in order to verify the validity of the model, the optimum extraction variables are obtained by ultrasonic glycyrrhizic acid extraction experiment, and on the basis of the optimum extraction variables, the content of glycyrrhizic acid and ultrasonic power, ultrasonic frequency and extraction temperature are established respectively. The dynamic model, experimental and simulation results show that the model is feasible and effective. (2) based on the support vector regression (SVR) theory, the prediction model of ultrasonic Glycyrrhiza extraction process based on support vector regression is established, which is based on the theory of support vector regression (SVR). The prediction results verify the model. Furthermore, on the basis of the analysis of least squares support vector machine (LSSVM) theory, the prediction model of ultrasonic Glycyrrhiza extraction process of least squares support vector machine (LS SVM) is constructed on the basis of the long training time of the support vector regression model. The model is compared with the prediction results of the support vector regression model (3). In order to optimize the parameter optimization problem in the least squares support vector machine modeling, a chaotic dynamic step size fruit fly optimization algorithm (CDSFOA) is proposed by introducing the dynamic step length regulation factor and chaos optimization algorithm. The algorithm converges to the global optimal solution by using the Markoff convergence analysis theory, and the algorithm is verified by example simulation. Based on this algorithm, the model parameters are optimized and the prediction model of ultrasonic Glycyrrhiza extraction process based on CDSFOA-LSSVM is established. The model is compared with the prediction results of support vector regression and least square support vector machine model, and the model has faster training speed and higher prediction precision. (4) (4) in view of the problem that the above model can not be predicted online, the online unbiased LSSVM algorithm is proposed with the unbiased LSSVM and online LSSVM as the theory. Based on the online unbiased LSSVM algorithm and the multi input and multi output LSSVM algorithm, a multi input and multi output online learning unbiasing is proposed. In order to further improve the computing speed and accuracy of the algorithm, the recursive formula of the algorithm is improved by introducing the weighting factor and using the online recursive learning method. Based on this, the ultrasonic multi target prediction model is established to predict the concentration of two effective components of rutin and quercetin in natural spear and guard. (5) in view of the problem that the current ultrasonic extraction process has a single frequency and can not be selected continuously, an ultrasonic frequency optimization method is proposed, that is to search the optimal frequency band in a wide range first in the broadband band, and then to search the optimal frequency in a small range in the optimal narrow band, and search continuously and automatically. The optimal ultrasonic frequency is used to improve the efficiency of ultrasonic natural extraction. Based on the ultrasonic extraction system developed by the frequency optimization idea, the effective components of two natural objects, edible tomato and Chinese Sophora japonica, were extracted and the frequency optimization was studied. The optimal ultrasonic frequencies were obtained. Better extraction efficiency was obtained.

【学位授予单位】：江南大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：Q946;TB559

【参考文献】