中文摘要：

采用电子鼻系统对秸秆饲料固态发酵过程阶段进行监测研究具有明显的应用意义。但是电子鼻系统传感器阵列中同一气敏传感器会对多种被测气体响应,导致采集数据含有冗余信息,因此有必要对电子鼻传感器阵列进行优化。本文基于因子分析法对电子鼻系统采集数据结合阶段状态信息进行分析,提出传感器阵列优化方法。并采用神经网络、支持向量机和高斯过程等模式识别方法对电子鼻系统传感器阵列优化组合采集数据进行过程状态识别模型建模。研究表明,传感器阵列优化有利于减少模型输入,降低模型复杂性,提高模型对过程状态的识别率。

英文摘要：

In this paper,e-nose system was attempted to monitor the state of wheat-straw solid-state fermentation（SSF）process in real time.In this e-nose system,the e-nose sensor array consists of eleven gas sensors in different types.When e-nose system is used to monitor the SSF process,each sensor in the system responds to various aromas,and the components of each aroma are in turn detected by many sensors.This causes high modeling complexity and low recognition rate.Therefore,it is necessary to optimize the e-nose sensor array.In the work,factor analysis（FA）method is used to reduce the dimensions of collected data by e-nose and select appropriate sensors to reduce the cost and the computation burden of pattern recognition methods.Neural network（NN）,support vector machine（SVM）and Gaussian process（GP）discrimination models were used to evaluate the e-nose array optimization method.And the overall results sufficiently demonstrate optimized e-nose array has higher identification performance than that of the original sensor array in identifying wheat-straw SSF stages.