Variable Correlation Analysis-Based Convolutional Neural Network for Far Topological Feature Extraction and Industrial Predictive Modeling

In process industries, accurate prediction of critical quality variables is particularly important for process control and optimization. Usually, soft sensors have been developed to estimate the quality variables via process variables. However, there are often process variables that are far apart in topology but have high correlations, presenting challenges for such feature learning in soft sensors. To solve these difficulties, a variable correlation analysis-based convolutional neural network (VCA-CNN) is proposed in this article for far topological feature extraction, which generates three-channel data samples containing natural topology features, linear topology features, and nonlinear topology features through Pearson and Spearman correlation to adequately focus on far-topology features. In the feature extraction stage, the VCA-CNN employs a special row convolution kernel to extract far-topological features. Then, normal convolutional and pooling operations are alternately processed on the output feature maps. This approach overcomes the limitation that traditional convolutional operations mainly extract local correlations of process variables. The effectiveness and superiority of the VCA-CNN are validated in the hydrocracking process and debutanizer column.

I. Introduction

In modern industries, the prediction of key quality variables and real-time monitoring are essential for enhancing productivity and ensuring stability. However, online measurement becomes challenging for key quality variables due to technological constraints, such as the lack of measuring equipment, the low reliability of analyzers, and significant measurement delays in most processes [1], [2]. As a solution, soft sensor technology is employed for quality variable prediction. Soft sensor technology utilizes predictive models based on easily measurable process variables to estimate quality variables that are difficult to measure directly [3], [4]. By leveraging soft sensor technology, the process industry can overcome the limitations of conventional measurement techniques and improve production efficiency and quality [5].