特征空间递归更新的ICA算法及发酵过程监测应用

及时更新监测模型以适应过程的时变特性，对准确检测出化工过程异常和设备故障具有重要意义。针对普通独立元分析（ICA）算法在更新计算监测模型时计算复杂度高、效率低的缺点，提出了一种基于特征空间递归更新的在线独立元分析（RUFS-ICA）算法。将算法应用于青霉素发酵过程的在线建模与监测中，与普通ICA方法相比，仿真统计结果表明，平均误警率降低至1.67%，基本克服了漏报现象；与其他在线更新算法相比，复杂度明显降低，计算时间减少54.1%，节省了存储量。

ICA algorithm based on recursively updating of feature space and application to fermentation process monitoring

Updating the monitoring model timely to address the time-variant characteristics has vital significance in detecting abnormalities of chemical process and equipment breakdown exactly.The conventional ICA-based methods used to update the model have a high computational load and low efficiency.RUFS-ICA algorithm was proposed to model and monitor a fed-batch penicillin fermentation process on line.Compared with conventional ICA methods, the proposed method reduced the false alarm rate to 1.67% and basically overcame alarm failures.Compared with other methods, the algorithm could greatly decrease the computation time by about 54.1%, and saved the memory.