Integrated framework of nonlinear prediction and process monitoring for complex biological processes

Abstract

Bioprocesses and biosystems have nonlinear and multiple operation patterns depending on the influent loads, temperatures, the activity of microorganisms, and other factors. In this paper, an integrated framework of nonlinear modeling and process monitoring methods is developed for a complex biological process. The proposed method is based on modeling by fuzzy partial least squares (FPLS) and on process monitoring by a statistical decomposition, which is suitable for predicting and supervising a nonlinear biological process. Case studies in the bio-simulated process and industrial biological plant show that the proposed method can give superior prediction and monitoring performance in complex biological plants compared to other linear and nonlinear methods, since it can effectively capture the nonlinear causal relationship within the biosystem. This gives us the integrated framework that is able to both model and monitor the nonlinear bioprocess simultaneously.

Challenges in the Development of Soft Sensors for Bioprocesses: A Critical Review

“…Another combination of the symptom signal and the contribution-based method was proposed by Yoo and Lee (2006) for sensor fault detection. Here, contribution plots assist in identifying faulty variables.…”

Dynamic Nonlinear Partial Least Squares Modeling Using Gaussian Process Regression

“…In recent years, process modeling has gained significant attention in chemical and biological processes that ensures the accurate prediction for key variables, such as product quality and pollutant emission. , As an essential part of multivariate statistical methods, latent variable methods (LVMs) are specifically suitable for coping with the high dimensionality and collinearity problems of industrial process data through the use of the latent variables that can be then used for constructing the regressions between the latent space and the output space . Among the different kinds of LVMs, partial least-squares (PLS) is the powerful one that takes both the variance structure and the correlation between the inputs and the outputs into consideration, and has been widely used for regression purposes. , However, when facing the obvious nonlinear characteristics of data, the modeling accuracy of the conventional linear PLS can be decreased significantly. , …”

“…5,6 However, when facing the obvious nonlinear characteristics of data, the modeling accuracy of the conventional linear PLS can be decreased significantly. 7,8 To handle the aforementioned problem, nonlinear extensions of PLS were proposed, and can be roughly divided into two categories. 9 The first one is the kernel-based PLS (KPLS), and the other one is a building of the nonlinear regression between each pair of latent variables of PLS.…”