Developing a Soft Sensor with Online Variable Selection for Industrial Multi-mode Processes

Abstract

Abstract Soft sensors are used to predict response variables, which are difficult to measure, using the data of predictors that can be obtained relatively easier. Arranging time-lagged data of predictors and applying partial least squares (PLS) to the dataset is a popular approach for extracting the correlation between data of the responses and predictors of the process dynamic. However, the model input dimension dramatically soars once multiple time delays are incorporated. Since irrelevant inputs deteriorate the prediction performance of the soft sensor, the selection of variables in the dynamic PLS (DPLS) model is a critical step for the robustness and the accuracy of the inferential model. In the literature, the soft sensors were updated using the newest operating data in order to track the time-varying characteristics of a real process. However, in industrial processes, the important variables may not be time invariant under the different operating conditions. In this study, an adaptive soft sensor is developed by incorporating with online variable reselection. The proposed approach is validated by the operating data from the industrial deethane process.