Online prediction of quality‐related variables for batch processes using a sequential phase partition method

Abstract

Batch processes inherently have multiple operation phases; different phases exhibit different characteristics. Hence, it is reasonable to partition the process into phases and build sub‐phase models for online quality prediction. To this end, a sequential phase partition method based on the information increment is proposed. To address the multiphase behaviours in batch processes, this work utilizes a new information increment index to capture the dynamic characteristics of batch processes along a time direction and divides the process into sub‐phases. Next, phase‐based multiway partial least squares (MPLS) models are built to model within‐phase characteristics and predict the quality‐related variables online. Information increment is able to exploit the process evolution by focusing on the changing variable correlations derived from two adjacent extend time slice. It directly utilizes the available process measurements of successful history batch processes without data transformation or dimensionality reduction. The method is sequential and can overcome the limits of some phase partition methods that may divide the samples with discontinuous time sequence but similar characteristics into the same phase. In addition, the information increment is capable of reflecting the change of the process intuitively with high computation efficiency. Advantages of the proposed method are illustrated by two case studies, a penicillin simulation platform and an industrial application of Escherichia coli (E. coli) fermentation, respectively.

Multihorizons transfer strategy for continuous online prediction of time‐series data in complex systems

“…where a b c τ , , , are the real numbers. When τ > 16.8, chaotic time series can be generated by formula (18). When generating source domain data, we set the parameter values to a b c τ = 0.2, = 0.1, = 10, = 17, and the target domain data are generated by changing the time step size.…”

“…13,14 Wang et al [15][16][17] studied the online sequence extreme learning mechanism with kernels (Os-ELMK), and made corresponding improvements to it by embedding the kernel or adding the forgetting mechanism, so as to improve its online learning performance on the original basis and obtained better accuracy. Li et al 18 proposed a sequential phase segmentation method based on information increment to solve the multistage problem in batch processing. This method captures the dynamic characteristics of batch processing along the time direction.…”

Soft Sensor Framework Based on Semisupervised Just-in-Time Relevance Vector Regression for Multiphase Batch Processes with Unlabeled Data

“…To address this issue, some phase-based methods have been developed. 13,14 These methods divide the whole batch process into multiple phases through phase partitioning methods and establish a model for each phase. However, the prediction performances of these methods depend on a large amount of labeled data.…”

“…They mainly focus on establishing a global regression model for the whole batch process, which ignores the multiphase characteristics of batch process data. To address this issue, some phase-based methods have been developed. , These methods divide the whole batch process into multiple phases through phase partitioning methods and establish a model for each phase. However, the prediction performances of these methods depend on a large amount of labeled data.…”

Soft sensor development based on improved just‐in‐time learning and relevant vector machine for batch processes

“…Soft sensors achieve online measurement of primary variables by establishing models between primary variables and process variables that are easy to measure online (also called secondary variables). [ 7,8 ] According to the modelling principle, there are two main types of soft sensors: the mechanism‐based soft sensor [ 9,10 ] and the data‐driven‐based soft sensor. [ 11 ] Compared with the mechanism‐based soft sensor, the data‐driven‐based soft sensor does not require prior process knowledge and is simple in application and highly flexible.…”