Batch Process Monitoring Based On Multi-way Independent Principal Component Analysis Methods

Batch processes have become more and more important in modern industrial processes. In ensuring the safety and stability of batch processes and high quality final product, on-line monitoring and fault diagnosis in batch processes emerge as an essential and important task. As the development of on-line measurement instruments and computer technology, large amounts of process variables'data can be collected more easily than before. The data can be analyzed to supervise the process behavior, by mining the valuable information and resources. The technology of fault diagnosis in the processes monitoring was an important research direction in the control field.In view of batch-to-batch variation in most industrial batch processes, an adaptive MPCA method is studied to capture the dynamic variation among different batches. This approach first establishes an MPCA model based on the historical database, then utilizes the initial monitoring model to detect the next new batch data. If the whole batch behavior is detected normally, we add the new normal batch to model database as the last batch and the oldest one is removed. Otherwise if some fault is detected in the new batch, the MICA model need not updated. The simulation results on a penicillin fermentation process show that the adaptive MPCA method can establish a more accurate statistical model, which can reduce false positives when detecting a normal batch. In order to keep early historical information, the adaptive MPCA method is improved. In the establishment of a new model, the forgetting factor is introduced in the improved method. The simulation results on a penicillin fermentation process show that the improved method can establish a more accurate statistical model, which can more effectively reduce false positives when detecting a normal batch.Data obtained from actual industrial system is very complex, which includes both Gaussian signal and non-Gaussian signal, so using PCA or ICA methods alone can not fully handle the data acquisition system. On this basis, we combine the advantages and disadvantages of them, the Multi-way Independent Principal Component Analysis (MIPCA) method is proposed which used to solve the above problems. The basic principle of the algorithm is that, the normal batch data is firstly pretreated, then maximize the use of negative entropy principle for data classification, the data in the non-Gaussian signal and Gaussian signal were modeled by MICA method and MPCA method, then on the need to detect the data in accordance with the above method of classification, I 2statistics, SPE statistical, T 2statistics and Q statistical methods are used to detect normal data. If test results are not the normal failure, contribution plans is used for fault diagnosis. The simulation in the penicillin fermentation process show that, detecting three kinds of typical fault, the results show the MIPCA methods have increased the sensitivity and accuracy.