Important process or product quality parameters in chemical plants are difficult to measure with sensors for economic or technical reasons and soft measurement is an important solution to measure these key parameters. Aiming at the strong nonlinearity, low prediction accuracy, frequent dynamic changes, and severe collinear interference in actual chemical production processes, this article proposes a dynamic soft sensor model using novel nonlinear autoregression with external input (NARX) based on principal component analysis (PCA) and wavelet denoising (WD) (PCA-WD-NARX). The feature information of the sample data is extracted by the PCA and the collinearity is eliminated at the same time. Moreover, the noise in the data set is eliminated by WD to simplify the complexity of the learning data. Then, the NARX is used to construct the dynamic soft sensor model. Finally, the recommended model is applied to predict the acetic acid content of a purified terephthalic acid (PTA) plant. Through the evaluation indexes of the root-mean-square error (RMSE) and the coefficient of determination (R 2), the experimental results show that the proposed method has the most outstanding prediction accuracy and generalization ability among the NARX model, the NARX integrating the PCA (PCA-NARX) model, the NARX integrating the WD (WD-NARX) model, the Elman model, and the recursive radial basis function (RRBF).

Polypropylene is an important raw material for producing medical masks. The melt index (MI) is one of the most important quality indexes in the propylene polymerization (PP) production process, but it cannot be physically measured in real time. In consideration of the strong nonlinearity, obvious dynamic characteristics, and complex mechanism of the PP process, the gray soft sensor model, which combines the merits of mechanism-driven modeling and data-driven modeling, has great research value. In this study, we propose a novel gray dynamic soft sensor modeling strategy. The influence factors of the MI are analyzed based on the process mechanism of PP production plants to select appropriate process variables and make necessary mechanism transformation. Then, the kernel principal component analysis and wavelet denoising are used to eliminate the multicollinearity and “noise” interference among process variables. Finally, an improved orthogonal sparse echo state network is used to construct the gray dynamic soft sensor model. The experimental results based on the real field data of the PP production plant show that the orthogonalization and sparseness of the reservoir can effectively enhance the performance of the reservoir and improve the operational efficiency. Meanwhile, the proposed dynamic soft sensing model has better prediction ability than the corresponding methods. Moreover, this study is of great significance to guide and optimize the PP production process.

No abstract