中文摘要：

在微生物发酵工业生产中，为保证产品的产量及质量．需要对与其相关的过程变量进行实时的监控；然而在实际过程中．直接与生产相关的生化参数往往无法或难以用传感器直接检测．只能通过采样离线人工化验的方法得到，因此针对实际生产需要并且根据发酵的工艺需求，提出一种基于BP神经网络的测量系统，系统采用BP神经网络与主元分析以及机理模型相结合的模型，弥补了单纯采用BP神经网络的缺陷；实验结果表明，在通过有效数据处理以及发酵动力学分析后，人工智能算法能够有效准确地实现在线的菌体检测。为控制算法提供数据，大大提高了产物浓度。

英文摘要：

In the microorganism fermentation industry production, it need to carry on the variable for the real--time monitoring to guarantee the product and the quality. However in the actual process, that is difficulty to monitor or hardly use the sensor to examining the production directly. So it can be obtained through the sampling off line manual chemical examination method. Therefore aims at the actual production need, and according to the fermentation craft demand, this article proposed a kind of the soft sensor system based on BP neural network. The experimental result indicated： After data processing as well as fermentation dynamics analysis, the system can accurate realize the online mycelium examination.