On-Line Monitoring of Fed-Batch Penicillin Cultivation Using Time-Varying and Multivariate Statistical Analysis

On-line monitoring of penicillin cultivation processes is crucial to the safe production of high-quality products. Multi-way principal component analysis(MPCA), a multivariate projection method, has been widely used to monitor batch and fed-batch processes. However, when MPCA is used for on-line batch monitoring, the future behavior of each new batch must be inferred up to the end of the batch operation at each time and the batch lengths must be equalized. This represents a major shortcoming because predicting the future observations without considering the dynamic relationships may distort the data information, leading to false alarms. In this paper, an improved statistical batch monitoring and fault diagnosing approach based on variable-wise unfolding is proposed, which include the following three aspect:(1)time-varying score covariance is used to replace fixed covariance at each time during a batch, which is considered to incorporate the major dynamic characteristics of the batch process and can obtain better monitoring performance, (2)principal-component-related variable residual statistic is introduced to replace SPE-statistic, which can avoid the conservation of SPE statistical test and provide more explicit information about the process conditions and (3)time-varying contribution charts is proposed to diagnose anomalous batch process. The proposed method was used to detect and identify faults in the fed-batch penicillin cultivation process, for four different scenarios. The simulation results clearly demonstrate the power and advantages of the proposed method in comparison to MPCA.