This paper discusses the design of an inferential sensor for the online prediction of the end-quality of an industrial batch polymerization process. Owing to unequal batch speeds, measurement profiles must be synchronized before modeling. This makes profile alignment an integral part of any inferential sensor. In this work, a novel online hybrid derivative dynamic time warping data alignment technique is presented. The proposed technique allows for automatic adjustment of the warping resolution to achieve optimal alignment results for both slowly and rapidly varying parts of the measurement profiles. The proposed online data alignment technique is combined with a multiway partial least-squares black box model to yield online predictions of the final quality of a running batch process. It is demonstrated that this inferential sensor is capable of accurately predicting the quality online for an industrial polymerization process, even when the production process is only halfway, that is, well before lab measurements become available. As a result of this early warning, batches violating the quality specifications can be corrected or even stopped. This leads to fewer off-spec batches, saves production time, lowers operational costs, and reduces waste material and energy.
a b s t r a c tThis work considers the application of classification algorithms for data-driven fault diagnosis of batch processes. A novel data selection methodology is proposed which enables online classification of detected disturbances without requiring the estimation of unknown (future) process behavior, as is the case in previously reported approaches.The proposed method is benchmarked in two case studies using the Pensim process model of Birol et al. (2002) implemented in RAYMOND. Both a simple k Nearest Neighbors (k-NN) and complex Least Squares Support Vector Machine (LS-SVM) are employed for classification to demonstrate the generic nature of the proposed approach. In addition, the influence of different data pretreatment methods on the classification performance is discussed, together with a motivation for selecting the correct pretreatment steps. Finally, the influence of the number of available training batches is studied.The results demonstrate that a good classification performance can be achieved with the proposed data selection method even with a low number of faulty training batches by exploiting knowledge on the nature of the to-be-diagnosed faults in the data pretreatment. This provides a proof of concept for classification-based batch diagnosis and demonstrates the importance of incorporating process insight in the construction of data-driven process monitoring and diagnosis tools.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.