基于SVM的葛根素提取软测量系统的设计

Design of a soft sensor system for Puerarin extraction based on SVM

针对超声波高效率提取植物根茎类药材活性成分含量难以实时在线检测的问题,采用基于粒子群算法和支持向量机相结合的软测量建模方法。通过对辅助变量的测量来建立软测量预测模型,从而估计和推断待测量变量的值,并利用单片机进行辅助变量的数据采集和处理、软测量模型的计算及建立和模型的硬件实现,从而实现了对超声波提取植物根茎类药材葛根素的在线测量。与目前采用紫外分光光度计分时离线取样进行检测相比,该方法克服了无法直接得到提取率测量值、离线测量工作量大等缺点,具有实时性好、测量精度较高(测量相对误差控制在3%以内)和响应速度快等特点。

Since the high efficiency detection of the active components by ultrasonic from medicinal plants is difficult, this paper adopts the method of a soft sensor model based on particle swarm optimization and support vector machine algorithm in order to estimate and infer the values of measured variables. Using a microcontroller unit for data of auxiliary variables collection and processing, the microcontroller can calculate and build the soft sensor model and the hardware realization of the model. Therefore, the on-line detection of the rate of extraction of medicinal plants by ultrasonic is im- plemented. The soft sensor system overcomes the shortcomings associated with the used method of off-line sample detec- tion by UV spectrometry, namely the large workload and inability to measure the extraction rate directly. The new system has the following advantageous features： high accuracy （relative errors within 3%）, good generality, fast response, and good real-time performance.