中文摘要：

采用无机盐培养基毕赤酵母发酵生产重组人溶菌酶，以人工配置的标准溶液作为建模集，利用ATR—FTIR结合PLS建立甘油和甲醇的预测模型，模型具有良好的稳定性和预测性，甘油预测模型的SEC和SEV分别为0．171和0．532g／L，甲醇预测模型的SEC和SEV分别为0．129和0．248g／L。在5L发酵罐中运用建立的模型预测发酵液中甘油和甲醇的浓度，预测标准误差（SEP）分别为1．09和0．86g／L，表明模型稳定，能有效指导实际生产。比较5L发酵罐中分别采用FTIR调控和溶氧（DO-stat）调控甲醇浓度下的比生长速率（p）和比生产速率（Qp），结果表明，FTIR调控下比生长速率和比生产速率均高于溶氧（DO-stat）调控。

英文摘要：

Mid-infrared FTIR spectroscopy is an efficient tool for the monitoring of bioprocesses. The use of Fourier transform mid-infrared spectroscopy （FTIR） combined with PLS to predict the concentrations of glycerol and methanol in fed-batch cultivations of Pichia pastoris in a chemical medium was investigated by using the synthetic solutions as calibration set. Generally, all models were robust and performed well, the calibration model allowed stand-rd errors of calibration （SEC） and standard errors of validation （SEV） for glycerol and methanol concentrations of 0. 171, 0. 532, 0. 129 and 0. 248 g/L, respectively. The results from 5 L Fed-batch experiments showed that FTIR data had a good agreement with control data, with standard error of prediction （SEP） of 1.09 and O. 86 g/L, respectively. Specific growth rate and specific production rate were compared between FTIR and DO-star to control methanol concentration. Results showed that FTIR could limit the methanol concentration under 2% , which could maintain a higher specific growth rate and specific production rate that was effective to enhance the production.