Variable-Scale Probabilistic Just-in-Time Learning for Soft Sensor Development with Missing Data

Abstract

Just-in-time learning (JITL) has been widely applied to data-driven modeling to deal with the nonlinearity problems in industrial processes. To mitigate the effects of noise existing in JITL, probabilistic JITL (PJITL) selects samples based on the probability distributions. Considering the existence of missing data situation, the PJITL algorithm could also cope with that. However, traditional JITL-based methods, including PJITL, cannot flexibly select the number of training samples for each query sample, which would in return influence the accuracy of prediction for a part of query samples. To solve this problem, we proposed a method named “variable-scale PJITL” (VS-PJITL) which can determine the sizes of the local models for each query sample using a new sample selection criterion. Based on the Euclidean distance, the sample selection criterion also applies to the variable-scale JITL (VS-JITL). Then, comparisons of VS-PJITL, PJITL, JITL, and VS-JITL are tested on a simulated data set and a real industrial data set from the catalytic naphtha reforming process. By analyzing the two cases above, VS-PJITL is considered to have superior performance to the original PJITL (root-mean-square error reduced by 0.3355 and 0.4778).

Deep Semi-Supervised Just-in-Time Learning Based Soft Sensor for Mooney Viscosity Estimation in Industrial Rubber Mixing Process

“…It mainly As can be seen from Figure 1, the definition of a similarity measure is the key to constructing high-performance JIT soft sensor models. The most commonly used similarity criteria include distance-based similarity functions, such as Euclidean distance [34,35] and Mahalanobis distance [36], angle-based similarity functions, such as cosine similarity [37,38], and similarity functions based on correlation coefficients, such as Pearson correlation coefficient similarity [39]. Among them, Euclidean-distance similarity is the simplest and most commonly used similarity criterion for JIT soft sensor modeling.…”

Distributed Missing Values Imputation Schemes for Plant-Wide Industrial Process Using Variational Bayesian Principal Component Analysis

“…Since most of the data-driven intelligent decision support systems are developed on complete data sets, the MVI technique is more universal. Recently, a number of MVI algorithms have been presented. − The mean imputation uses the average value of a variable’s observations to impute the variable’s MV, but it results in a smaller variance . The k nearest-neighbor (kNN) imputation is to find out the k nearest-neighbor samples from the data set except for the samples that have MV at the same positions as those to be imputed .…”

Study on Characterization and Coke Compositions of Deactivated 5 A Molecular Sieve for Adsorption Separation of Industrial Naphtha

“…Naphtha is a light oil composed of C 6 −C 10 linear/branched alkanes, naphthenes or aromatics. It is an important raw material for cracking to produce ethylene and propylene, catalytic reforming to produce benzene, toluene and xylene, as well as high‐octane blending material of gasoline [1] . Same carbon number n‐paraffins/isoparaffins in the naphtha have near boiling point and relative volatility close to 1, and they are difficult to effectively separate by an ordinary distillation process.…”

“…It is an important raw material for cracking to produce ethylene and propylene, catalytic reforming to produce benzene, toluene and xylene, as well as high-octane blending material of gasoline. [1] Same carbon number n-paraffins/isoparaffins in the naphtha have near boiling point and relative volatility close to 1, and they are difficult to effectively separate by an ordinary distillation process. The adsorption separation technology can achieve an efficient separation of n-paraffins/isoparaffins in the naphtha through a molecular scale, and the product with rich nparaffins can be used as the cracking raw materials for ethylene, the product containing rich iso-paraffins can be used as the catalytic reforming raw materials or high-octane blending materials of gasoline.…”